WO2021143241A1 - 图像处理方法、装置、电子设备及存储介质 - Google Patents

图像处理方法、装置、电子设备及存储介质 Download PDF

Info

Publication number
WO2021143241A1
WO2021143241A1 PCT/CN2020/122203 CN2020122203W WO2021143241A1 WO 2021143241 A1 WO2021143241 A1 WO 2021143241A1 CN 2020122203 W CN2020122203 W CN 2020122203W WO 2021143241 A1 WO2021143241 A1 WO 2021143241A1
Authority
WO
WIPO (PCT)
Prior art keywords
color feature
histogram
interval
feature interval
color
Prior art date
Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
Ceased
Application number
PCT/CN2020/122203
Other languages
English (en)
French (fr)
Inventor
李志成
Current Assignee (The listed assignees may be inaccurate. Google has not performed a legal analysis and makes no representation or warranty as to the accuracy of the list.)
Tencent Technology Shenzhen Co Ltd
Original Assignee
Tencent Technology Shenzhen Co Ltd
Priority date (The priority date is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the date listed.)
Filing date
Publication date
Application filed by Tencent Technology Shenzhen Co Ltd filed Critical Tencent Technology Shenzhen Co Ltd
Priority to EP20913353.7A priority Critical patent/EP4024327A4/en
Priority to KR1020227006382A priority patent/KR102617626B1/ko
Priority to JP2022520267A priority patent/JP7508135B2/ja
Publication of WO2021143241A1 publication Critical patent/WO2021143241A1/zh
Priority to US17/573,491 priority patent/US12260521B2/en
Anticipated expiration legal-status Critical
Ceased legal-status Critical Current

Links

Images

Classifications

    • G—PHYSICS
    • G06—COMPUTING OR CALCULATING; COUNTING
    • G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
    • G06T5/00—Image enhancement or restoration
    • G06T5/40—Image enhancement or restoration using histogram techniques
    • G—PHYSICS
    • G06—COMPUTING OR CALCULATING; COUNTING
    • G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
    • G06T3/00—Geometric image transformations in the plane of the image
    • G06T3/40—Scaling of whole images or parts thereof, e.g. expanding or contracting
    • G06T3/4053—Scaling of whole images or parts thereof, e.g. expanding or contracting based on super-resolution, i.e. the output image resolution being higher than the sensor resolution
    • G—PHYSICS
    • G06—COMPUTING OR CALCULATING; COUNTING
    • G06F—ELECTRIC DIGITAL DATA PROCESSING
    • G06F18/00—Pattern recognition
    • G06F18/20—Analysing
    • G06F18/22—Matching criteria, e.g. proximity measures
    • G—PHYSICS
    • G06—COMPUTING OR CALCULATING; COUNTING
    • G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
    • G06T3/00—Geometric image transformations in the plane of the image
    • G06T3/40—Scaling of whole images or parts thereof, e.g. expanding or contracting
    • G—PHYSICS
    • G06—COMPUTING OR CALCULATING; COUNTING
    • G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
    • G06T5/00—Image enhancement or restoration
    • G06T5/70—Denoising; Smoothing
    • G—PHYSICS
    • G06—COMPUTING OR CALCULATING; COUNTING
    • G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
    • G06T7/00—Image analysis
    • G06T7/90—Determination of colour characteristics
    • G—PHYSICS
    • G06—COMPUTING OR CALCULATING; COUNTING
    • G06V—IMAGE OR VIDEO RECOGNITION OR UNDERSTANDING
    • G06V10/00—Arrangements for image or video recognition or understanding
    • G06V10/40—Extraction of image or video features
    • G06V10/50—Extraction of image or video features by performing operations within image blocks; by using histograms, e.g. histogram of oriented gradients [HoG]; by summing image-intensity values; Projection analysis
    • G—PHYSICS
    • G06—COMPUTING OR CALCULATING; COUNTING
    • G06V—IMAGE OR VIDEO RECOGNITION OR UNDERSTANDING
    • G06V10/00—Arrangements for image or video recognition or understanding
    • G06V10/40—Extraction of image or video features
    • G06V10/56—Extraction of image or video features relating to colour
    • H—ELECTRICITY
    • H04—ELECTRIC COMMUNICATION TECHNIQUE
    • H04N—PICTORIAL COMMUNICATION, e.g. TELEVISION
    • H04N1/00—Scanning, transmission or reproduction of documents or the like, e.g. facsimile transmission; Details thereof
    • H04N1/40—Picture signal circuits
    • H04N1/40068—Modification of image resolution, i.e. determining the values of picture elements at new relative positions
    • H—ELECTRICITY
    • H04—ELECTRIC COMMUNICATION TECHNIQUE
    • H04N—PICTORIAL COMMUNICATION, e.g. TELEVISION
    • H04N1/00—Scanning, transmission or reproduction of documents or the like, e.g. facsimile transmission; Details thereof
    • H04N1/46—Colour picture communication systems
    • H04N1/56—Processing of colour picture signals
    • H04N1/58—Edge or detail enhancement; Noise or error suppression, e.g. colour misregistration correction
    • G—PHYSICS
    • G06—COMPUTING OR CALCULATING; COUNTING
    • G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
    • G06T2207/00—Indexing scheme for image analysis or image enhancement
    • G06T2207/10—Image acquisition modality
    • G06T2207/10024—Color image

Definitions

  • This application relates to the field of computer technology, and in particular to an image processing method, device, electronic equipment, and storage medium.
  • the existing up-sampling algorithms mainly include interpolation algorithms or algorithms based on deep learning.
  • the basic principle is to calculate the color feature values of inserted pixels based on the color feature values of adjacent pixels, because these algorithms only rely on local adjacent pixels. To calculate the color information of the inserted pixels, the up-sampled image will appear obvious jagged and color edge noise, which reduces the quality and display effect of the up-sampled image.
  • An image processing method including:
  • the color characteristic values of the pixel points in the up-sampled image falling in the respective color characteristic intervals are processed to obtain the target image.
  • An image processing device including:
  • the first statistics module is configured to obtain a first histogram corresponding to the original image according to the number of pixels in the original image whose color feature values fall within each color feature interval;
  • a second statistics module configured to obtain a second histogram corresponding to the up-sampled image according to the number of pixels in the up-sampled image of the original image whose color feature values fall within the respective color feature intervals;
  • a matching module configured to respectively determine the target color feature interval matched by each color feature interval in the second histogram according to the first histogram and the second histogram;
  • the processing module is configured to process the color feature values of pixels in the up-sampled image that fall into the respective color feature intervals according to the target color feature intervals matched by the respective color feature intervals in the second histogram, Obtain the target image.
  • An electronic device includes a memory and one or more processors.
  • the memory stores computer readable instructions.
  • the one or more processors execute the steps of the image processing method.
  • One or more non-volatile computer-readable storage media storing computer-readable instructions.
  • the computer-readable instructions are executed by one or more processors, the one or more processors execute the steps of the above-mentioned image processing method.
  • FIG. 1A is a schematic diagram of an application scenario of an image processing method provided by an embodiment of this application.
  • FIG. 1B is a schematic diagram of an application scenario of an image processing method provided by another embodiment of this application.
  • FIG. 1C is a schematic diagram of an image processing method provided by an embodiment of the application applied to a neural network
  • FIG. 1D is a schematic diagram of an image processing method provided by another embodiment of the application applied to a neural network
  • FIG. 2 is a schematic flowchart of an image processing method provided by an embodiment of the application.
  • FIG. 3 is a gray value corresponding to an image provided by an embodiment of the application and a histogram obtained by statistical gray value;
  • FIG. 4 is a schematic flowchart of determining the target color feature interval matched by each color feature interval in the second histogram according to an embodiment of the application;
  • FIG. 5 is a corresponding diagram of an up-sampled image and a target image provided by an embodiment of the application
  • FIG. 6 is a schematic structural diagram of an image processing device provided by an embodiment of the application.
  • FIG. 7 is a schematic structural diagram of an electronic device provided by an embodiment of the application.
  • Histogram also known as quality distribution chart, is a statistical report chart.
  • a series of vertical stripes or line segments of varying heights represent the data distribution.
  • the horizontal axis represents the data type and the vertical axis represents the distribution.
  • the color feature value is divided into multiple color feature intervals in advance, that is, the horizontal axis of the histogram is the color feature interval, and the number of color feature values of pixels in an image falling into each color feature interval is counted. Obtain a histogram that characterizes the color feature distribution of the image.
  • RGB It is a color standard in the industry. Various colors are obtained by changing the three color channels of red (R), green (G), and blue (B) and superimposing them with each other. RGB is the color representing the three channels of red, green, and blue. This standard includes almost all colors that human vision can perceive, and it is one of the most widely used color systems.
  • YUV It is a color coding method that is often used in various video processing components. When YUV encodes photos or videos, it takes into account human perception and allows the bandwidth of chroma to be reduced.
  • YUV is a type of compiling true-color color space (color space).
  • Y'UV, YUV, YCbCr, YPbPr, etc. can all be called YUV, which overlap with each other.
  • "Y” represents brightness (Luminance or Luma), which is the grayscale value
  • "U” and "V” represent chrominance (Chrominance or Chroma), which are used to describe the color and saturation of the image, and are used to specify pixels s color.
  • CIELab is a color system of CIE, the color system, based on CIELab means based on this color system, basically used to determine the numerical information of a certain color.
  • Lab mode is a color mode published by the International Commission on Illumination (CIE) in 1976. It is a color model determined by the CIE organization that theoretically includes all colors visible to the human eye. Lab mode makes up for the deficiencies of the two color modes of RGB and CMYK. It is an internal color mode used by Photoshop to convert from one color mode to another. Lab mode is also composed of three channels, the first channel is lightness, that is, "L". The color of the a channel is from red to dark green; the color of the b channel is from blue to yellow.
  • the Lab mode In terms of expressing color range, the Lab mode is the most comprehensive, followed by the RGB mode, and the narrowest is the CMYK mode. That is to say, Lab mode defines the most colors, and has nothing to do with light and equipment, and the processing speed is as fast as RGB mode, and several times faster than CMYK mode.
  • Color feature interval multiple interval ranges obtained by dividing the range of color feature values corresponding to the color feature.
  • the color feature may be data indicating the color feature of the pixel such as the grayscale, brightness, color of the pixel.
  • the gray scale value range is 0 to 255
  • the gray scale value range from 0 to 255 can be divided into multiple interval ranges, and each interval range is used as a color feature interval.
  • 0-15 is a color characteristic interval
  • 16-31 is a color characteristic interval
  • so on a total of 16 continuous color characteristic intervals b1 ⁇ b16 can be obtained.
  • the level of the color characteristic interval bi can be recorded as i.
  • Subsampled It can also be called downsampled. Its purpose is to reduce the resolution of the image, that is, to reduce the number of pixels of the image so that the image fits the size of the display area or generate a thumbnail corresponding to the original image.
  • Upsampling It can also be Image Super Resolution. Its purpose is to improve the resolution of the image, that is, increase the number of pixels in the image, so that the image can adapt to high-resolution applications or restore the original image Lost details.
  • images can be up-sampled through image interpolation or algorithms based on deep learning. Common interpolation algorithms include: nearest neighbor interpolation, bilinear interpolation, bicubic interpolation, mean interpolation, median interpolation and other methods.
  • the original image can be up-sampled to restore or reconstruct more details in the original image, and then display it on a higher-resolution display device.
  • the input image is usually down-sampled and then input to the neural network, and then the output image of the neural network is uploaded.
  • At least one down-sampling network layer can also be set in the neural network, and the processing efficiency can be improved by reducing the resolution of the image during the processing, and by setting at least one up-sampling The network layer restores the resolution of the down-sampled image.
  • the relevant upsampling algorithms mainly include image interpolation algorithms or algorithms based on deep learning, etc.
  • the basic principle of any interpolation algorithm is to calculate the color characteristics of the inserted pixels based on the color characteristics of adjacent pixels.
  • the basic principle of the algorithm based on deep learning is to learn the relationship between neighboring pixels through the neural network to obtain the inserted pixels, because these algorithms only calculate the color information of the inserted pixels based on the color information of the local neighboring pixels , Resulting in obvious aliasing and color edge noise in the up-sampled image, and the adjacent pixel set in one or two directions of the top, bottom, left, and right will appear at the edge of the image.
  • the current algorithm can only be used Filling adjacent pixels causes the effect of pixels at the edge of the image to differ greatly from expected, which reduces the quality and display effect of the up-sampled image.
  • this application proposes an image processing method, which includes: obtaining the first histogram corresponding to the original image according to the number of pixel points in the original image whose color feature values fall within each color feature interval; The number of pixels in the up-sampled image whose color feature values fall within each color feature interval is obtained, and the second histogram corresponding to the up-sampled image is obtained; the second histogram is determined separately according to the first histogram and the second histogram The target color feature interval matched by each color feature interval in each color feature interval; According to the target color feature interval matched by each color feature interval, the color feature value of the pixel points in each color feature interval in the upsampled image is processed to obtain the target image .
  • the first histogram represents the histogram of the color feature distribution in the original image
  • the second histogram represents the histogram of the color feature distribution of the up-sampled image.
  • the target color feature interval matched by each color feature interval in the second histogram is determined, and then according to the matched target color feature interval, the color feature values of the pixels in the up-sampled image are determined one by one Make adjustments to obtain the target image corresponding to the up-sampled image, so that the color feature distribution of the target image corresponding to the up-sampled image is as consistent as possible with the color feature distribution of the original image.
  • This method can significantly reduce the aliasing in the up-sampled image , Color edge noise, etc., improve the quality and display effect of up-sampled images, better restore image details, and the processing method is simple and efficient, especially suitable for applications that require high processing efficiency, such as real-time video transmission.
  • FIG. 1A is a schematic diagram of an application scenario of an image processing method provided by an embodiment of the application.
  • This application scenario includes multiple terminal devices, including terminal device 101-1, terminal device 101-2, ... terminal device 101-n), and background server 102.
  • each terminal device and the back-end server 102 are connected via a wireless or wired network.
  • the terminal devices include but are not limited to desktop computers, smart phones, mobile computers, tablet computers, media players, smart wearable devices, televisions, surveillance cameras, etc.
  • the back-end server can be a server, a server cluster composed of several servers, or a cloud computing center.
  • the terminal device can upload images or videos to the backend server 102, or obtain images or videos from the backend server 102.
  • low-resolution images or videos can be transmitted, and then the low-resolution images or videos can be converted into high-resolution images or videos at the receiving end.
  • the terminal equipment Or the back-end server 102 can perform enhancement processing on the up-sampled image in combination with the image processing method provided in the present application when performing up-sampling processing on the image or video. Specific application scenarios are not limited to live video, video transmission, image transmission, etc.
  • a surveillance camera will down-sample the collected video and then transmit it to a monitoring display, and the monitoring display will perform up-sampling processing on the down-sampled video to obtain high resolution.
  • Video for display will be directly transmit images or video between two terminal devices.
  • a surveillance camera will down-sample the collected video and then transmit it to a monitoring display, and the monitoring display will perform up-sampling processing on the down-sampled video to obtain high resolution.
  • Video for display may directly transmit images or video between two terminal devices.
  • FIG. 1B is a schematic diagram of an application scenario of an image processing method provided by an embodiment of the application.
  • an image processing application is installed in the terminal device 111, and a software function module corresponding to the image processing method provided in this application is embedded in the image application.
  • the image processing method provided in this application can be based on the histogram of the original image and the histogram of the up-sampled image.
  • the up-sampled image is processed to obtain a high-resolution image.
  • the terminal device 111 includes, but is not limited to, a desktop computer, a mobile computer, a tablet computer, a smart phone, and other devices that can run image processing applications.
  • the image processing method provided in this application can also be embedded in image viewing software or image editing software.
  • image viewing software or image editing software When the user is viewing the image through the image viewing software or image editing software, if the user needs to focus on some details of the image, the image can be zoomed and transformed.
  • the image viewing software or image editing software can be based on the histogram of the original image
  • the graph and the histogram of the up-sampled image are processed to obtain a high-resolution image.
  • the image processing method provided in this application can be widely used in the processing of special images such as radar images, satellite remote sensing images, astronomical observation images, geological prospecting data images, biomedical slices, microscopic images, daily human scene images, and videos.
  • special images such as radar images, satellite remote sensing images, astronomical observation images, geological prospecting data images, biomedical slices, microscopic images, daily human scene images, and videos.
  • the image processing method provided in this application can also be applied to neural networks.
  • the input image is usually down-sampled and then input to the neural network, and then the neural network
  • the output image is up-sampled to restore the output image to the resolution of the input image.
  • the image processing method provided in this application can be used to enhance the up-sampled image to obtain the target image to improve the quality and display effect of the target image.
  • the neural network may also include at least one down-sampling process.
  • the processing efficiency is improved by reducing the resolution of the image during the process, and the resolution of the image will be restored through at least one up-sampling process.
  • an enhanced processing layer can be added after the up-sampling network layer, so as to use the image processing method provided in this application to process the up-sampling processed image to improve the quality of the up-sampling image, and then input the processed image to the next Network layer for processing.
  • an image processing method provided by an embodiment of the present application can be applied to a terminal device or a back-end server in the above-mentioned application scenario, and specifically includes the following steps:
  • S201 Obtain a first histogram corresponding to the original image according to the number of pixels in the original image whose color feature values fall within each color feature interval.
  • the original image may refer to the image before the upsampling process, and the original image may be a single image, or it may be an image frame by frame in the video.
  • the color feature in the embodiments of the present application refers to the feature that characterizes the color characteristics of the pixel, for example, the grayscale, brightness, color, etc. of the pixel.
  • the color characteristic value includes at least one of the following: the gray value of the pixel, the brightness value of the pixel, the color value of the pixel, and the like.
  • the color value can be determined according to the color system adopted by the image.
  • the color feature of each pixel in the image may include a grayscale value.
  • a histogram corresponding to the grayscale value can be obtained for one image.
  • the color characteristics of each pixel in the image include three color characteristics of R (red), G (green), and B (blue). These three color characteristics can be counted separately, namely For one image, three histograms corresponding to the three color features of R (red), G (green), and B (blue) can be obtained.
  • the color feature interval can be preset by those skilled in the art according to specific application scenarios and the range of feature values corresponding to the color feature, and a color feature interval can correspond to a color feature value or a range of color feature values.
  • the color feature is the grayscale of the pixel, and the grayscale value ranges from 0 to 255.
  • Each grayscale value can be regarded as a color feature interval, that is, a total of 256 grayscale intervals can be obtained; or 0 to 255
  • the gray value range of is divided into multiple regions, and each region is used as a color feature interval.
  • 0-15 is a color feature interval
  • 16-31 is a color feature interval
  • a total of 16 color feature intervals can be obtained b1 ⁇ b16, refer to Figure 3, which shows the gray value of each pixel in an image. For this image, count the number of pixels whose gray value falls within the 16 color feature intervals mentioned above to obtain the image The histogram corresponding to the gray value.
  • S202 Obtain a second histogram corresponding to the up-sampled image according to the number of pixels in the up-sampled image of the original image whose color feature values fall within each color feature interval.
  • the up-sampled image in the embodiment of this application is an image obtained after up-sampling the original image, and this application does not limit the specific up-sampling processing method.
  • S203 According to the first histogram and the second histogram, respectively determine the target color feature interval matched by each color feature interval in the second histogram.
  • the first histogram represents the histogram of the color feature distribution in the original image
  • the second histogram represents the histogram of the color feature distribution of the up-sampled image
  • the matching between each color feature interval in the second histogram and the color feature interval in the first histogram can be determined
  • the matching relationship reveals the corresponding relationship between the color feature intervals with the closest color feature distribution in the first histogram and the second histogram. Based on the matching relationship, it is determined that each color feature interval in the second histogram matches The target color feature interval.
  • the color feature values of the pixels in the up-sampled image are adjusted one by one to obtain the target image corresponding to the up-sampled image, so that the up-sampled image corresponds to The color feature distribution of the target image is as consistent as possible with the color feature distribution of the original image.
  • multiple color features can be adjusted at the same time.
  • the first histogram of the original image for each color feature is obtained .
  • the second histogram for each color feature of the up-sampled image and then, based on the first histogram and the second histogram corresponding to each color feature, the color feature value of each color feature in the up-sampled image Make adjustments.
  • the up-sampled image it is necessary to adjust the brightness and gray scale of the up-sampled image, and then obtain the first histogram VL1 corresponding to the brightness of the original image, the first histogram VG1 corresponding to the gray level of the original image, and the brightness of the up-sampled image The corresponding second histogram VL2, the second histogram VG2 corresponding to the gray level of the up-sampled image.
  • the up-sampled image P1 The brightness values of the pixels falling in each brightness interval are adjusted to the brightness values corresponding to the target brightness interval matched by each brightness interval, thereby obtaining the image P2.
  • the target gray-scale intervals matched by each gray-scale interval in the second histogram are respectively determined, and according to the target gray-scale intervals matched by each gray-scale interval, the The grayscale values of the pixels in the image P2 that fall into the grayscale intervals are adjusted to the grayscale values corresponding to the target grayscale intervals matched by the grayscale intervals, so as to obtain the target image P3.
  • the first histogram corresponding to the color features Y, U, and V of the original image, and the second histogram corresponding to the color features Y, U, and V of the up-sampled image, respectively, can be obtained.
  • the target color feature intervals matched by each color feature interval in the second histogram corresponding to the color feature Y are respectively determined, and according to the matching of each color feature interval
  • the value of the color feature Y of the pixel points in each color feature interval of the second histogram corresponding to the color feature Y in the up-sampled image P1 is processed to obtain the image P2.
  • the target color feature interval matched by each color feature interval in the second histogram corresponding to the color feature U is determined, and the target color feature interval is determined according to the color feature interval.
  • the value of the color feature U of the pixel points in each color feature interval of the second histogram corresponding to the color feature U in the image P2 is processed to obtain the image P3.
  • the target color feature interval matched by each color feature interval in the second histogram corresponding to the color feature V is determined respectively, and the target color feature interval is determined according to each color feature interval.
  • the value of the color feature V of the pixel points in each color feature interval of the second histogram corresponding to the color feature V in the image P3 is processed to obtain the target image.
  • the image processing method provided by the embodiments of the application can significantly weaken the jaggedness, color edge noise, etc. that appear in the up-sampled image, improve the quality and display effect of the up-sampled image, and better restore the image details, and the processing method is simple and efficient, especially It is suitable for applications that require high processing efficiency, such as real-time video transmission.
  • the matching relationship between each color feature interval in the second histogram and the color feature interval in the first histogram can be determined in a variety of ways, so as to determine the location of each color feature interval in the second histogram.
  • the matching target color feature interval can be determined in a variety of ways, so as to determine the location of each color feature interval in the second histogram.
  • the target color feature interval matched by each color feature interval in the second histogram may be determined according to the proportion values corresponding to the respective color feature intervals counted by the first histogram and the second histogram.
  • the proportion corresponding to each color feature interval in the first histogram is: the ratio of the number of pixels in the original image whose color feature values fall within the corresponding color feature interval to the total number of pixels contained in the original image
  • the proportion corresponding to each color feature interval in the second histogram is: the ratio of the number of pixel points in the up-sampled image whose color feature value falls within the corresponding color feature interval to the total number of pixels contained in the up-sampled image.
  • step S203 specifically includes: for any color feature interval in the second histogram, determining the difference between the proportion value corresponding to each color feature interval in the first histogram and the proportion value corresponding to any color feature interval Value, determine the target color feature interval matched by any color feature interval from the color feature interval corresponding to the difference value in the first histogram that satisfies the specified condition.
  • the specified condition may be the minimum value among all the differences corresponding to any color feature interval in the second histogram.
  • the first histogram and the second histogram both contain 16 color feature intervals b1 to b16.
  • the difference between the ratio and the proportion value corresponding to each color feature interval in the first histogram a total of 16 differences can be obtained, and the smallest difference is selected from these 16 differences, and the difference is in
  • the corresponding color feature interval in the first histogram is determined as the target color feature interval matched by the color feature interval b1 in the second histogram.
  • the smallest difference is the proportion of the color feature interval b1 in the second histogram
  • the difference between the value of the proportion of the color feature interval b3 and the color feature interval b3 in the first histogram, the target color feature interval matched by the color feature interval b1 in the second histogram is b3.
  • the specified condition can be the top N differences in the order, where N It is greater than or equal to 1 and less than the total number of color feature intervals.
  • the specific value of N can be determined by those skilled in the art according to actual application scenarios, and is not limited in the embodiment of the present application.
  • the color feature interval b1 in the second histogram calculates the colors in the second histogram respectively The difference between the proportion value of the characteristic interval b1 and the proportion value corresponding to each color characteristic interval in the first histogram, a total of 16 differences can be obtained, and the 16 differences are sorted in ascending order , Select the top 3 differences in sorting, assuming that the color feature intervals corresponding to the top 3 differences in the first histogram are b1, b3, b10, then determine one of b1, b3, and b10 The color feature interval is used as the target color feature interval matched by the color feature interval b1 in the second histogram.
  • the specified condition can also be that the difference is less than the difference threshold, that is, the difference less than the difference threshold corresponds to the color feature interval in the first histogram, and it is determined as the difference corresponding to the second histogram.
  • the target color feature interval matched by the color feature interval.
  • the difference between the proportion value of the color feature interval b1 in the second histogram and the proportion value of the color feature interval b4 in the first histogram is less than the difference threshold
  • the difference between the proportion value of the color feature interval b1 in the second histogram and the proportion value of the color feature interval b2 in the first histogram is less than the difference threshold, then one of the color feature intervals b2 and b4 is determined
  • the color feature interval is used as the target color feature interval matched by the color feature interval b1 in the second histogram.
  • the image processing method provided by the foregoing embodiments can simply and efficiently determine the matching of each color feature interval in the second histogram based on the difference in the proportion values between the respective color feature intervals in the first histogram and the second histogram
  • the target color feature interval can significantly weaken the jaggedness and color edge noise in the up-sampled image, improve the quality and display effect of the up-sampled image, and better restore the image details.
  • any color feature interval in the second histogram if there is only one difference that satisfies the specified condition, then the difference that satisfies the specified condition is determined as the corresponding color feature interval in the first histogram
  • the gray value for example 0-15 is a color feature interval b1 corresponding to the gray value range of 0-15
  • the color feature interval b2 corresponds to the gray value range of 16 to 31
  • the color feature interval b16 corresponds to the gray scale
  • the value range is 240-255.
  • the gray value between the color characteristic interval b1 and the color characteristic interval b2 is closer. Since multiple color feature intervals may have the same proportion value in the same image, for one color feature interval in the second histogram, there may be multiple color feature intervals in the first histogram with the same proportion value.
  • the proportion values of the color feature intervals in the second histogram are the same or similar, and the similar color feature intervals are the desired adjusted target color feature intervals.
  • the color feature interval b1 in the second histogram matched by the color feature interval in the first histogram is b2 and b5, respectively
  • the level difference between the color feature interval b2 and the color feature interval b1 is smaller than the level difference between the color feature interval b5 and the color feature interval b1, that is, the color feature value corresponding to the color feature interval b2 is closer to that corresponding to the color feature interval b1
  • the target color feature interval matched by the color feature interval b1 in the second histogram is b2.
  • the first histogram can be determined by some algorithms based on the distribution of the color feature values obtained by the statistics of the first histogram and the second histogram Based on the mapping relationship between the color feature interval and the color feature interval in the second histogram, the target color feature interval matched by each color feature interval in the second histogram can be determined based on the mapping relationship.
  • mapping relationship between the color feature intervals in the first histogram and the second histogram can be determined by the following formula:
  • MIN is the function to find the minimum value
  • Index is the function to obtain the color feature interval bj corresponding to the minimum value output by the function MIN
  • i represents the i-th color feature interval bi in the second histogram V2
  • V2[i] represents The proportion value of the i-th color feature interval bi in the second histogram V2
  • V1[j] represents the proportion value of the j-th color feature interval bj in the first histogram V1, 1 ⁇ j ⁇ n
  • n is the color feature The total number of intervals.
  • a MSE/618, where the mean square error
  • the parameter a can characterize the overall deviation of the proportion value of the color feature interval between the first histogram and the second histogram
  • the parameter c
  • increasing the parameter c can increase the probability that the color feature interval close to the color feature interval i in the second histogram is hit.
  • mapping relationship F(i) is not limited to the above-listed formulas, and the parameters a and c are not limited to the above-listed ways.
  • mapping relationship can also be:
  • MAX is the function of seeking the maximum value, guarantee F(i) ⁇ 1.
  • the first histogram and the second histogram are based on the difference in the proportion values between the respective color feature intervals, and the level difference between the color feature intervals is combined to determine the first histogram and the second histogram.
  • the target color feature interval matched by each color feature interval in the two histogram further improves the matching accuracy.
  • the target color feature interval matched by each color feature interval in the second histogram can be calculated according to the first histogram of the original image and the second histogram of the up-sampled image.
  • the first histogram counts the number of pixels in the original image whose color feature values fall within each color feature interval
  • the second histogram counts the color feature values in the up-sampled image fall within each color feature interval The number of pixels.
  • step S203 may include the following steps:
  • S401 Determine the first mean value and the first variance corresponding to the color feature interval of the pixel in the original image according to the first histogram, and determine the second corresponding to the color feature interval of the pixel in the up-sampled image according to the second histogram. Mean and second variance.
  • S402 Determine a first mapping relationship between the color feature interval in the first histogram and the color feature interval in the second histogram according to the first average value, the first variance, the second average value, and the second variance.
  • the first mapping relationship may be:
  • ⁇ D2/D1
  • round is a function of rounding to an integer
  • V2(i) is the number of pixels falling within the color feature interval bi in the second histogram
  • the target color feature interval matched by b1 is b2.
  • the first mapping relationship may also be:
  • MID is a function to find round( ⁇ (V2[i]-E2)+E1), the middle value of the three numbers 1 and n, and ensure that 1 ⁇ F(i) ⁇ n, round A function to find an integer for rounding.
  • the target color feature interval matched by b2 is b4, and the target color feature interval matched by each color feature interval in the second histogram is obtained by analogy.
  • the image processing method provided by the foregoing embodiment directly calculates the difference between the color feature intervals in the first histogram and the second histogram based on the proportion values between the color feature intervals in the first histogram and the second histogram. According to the mapping relationship, the target color feature interval matched by each color feature interval in the second histogram is determined according to the mapping relationship, and the processing method is simple and efficient.
  • step S204 specifically includes: if each color feature interval corresponds to a color feature value, adjusting the color feature values of the pixels in each color feature interval in the up-sampled image to The color feature value corresponding to the target color feature interval matched by each color feature interval obtains the target image.
  • each grayscale value is regarded as a color feature interval, that is, a total of 256 grayscale intervals b1 to b256 can be obtained. That is, each grayscale interval corresponds to a grayscale value.
  • the grayscale value 0 of the pixels falling in the grayscale interval b1 in the up-sampled image is changed to that corresponding to the target grayscale interval b2 Grayscale value 1; the target grayscale interval matched by the grayscale interval b2 in the second histogram is b2, so there is no need to adjust the grayscale value of the pixels falling in the grayscale interval b2 in the upsampled image.
  • step S204 specifically includes: if each color feature interval corresponds to a color feature value range, respectively determining the color feature value range and each color feature corresponding to each color feature interval in the second histogram The second mapping relationship between the color feature value ranges corresponding to the target color feature interval matched by the interval, the color feature value of the pixel points in each color feature interval in the up-sampled image is adjusted to the color determined according to the second mapping relationship Eigenvalues.
  • the grayscale value range is 0 ⁇ 255
  • the grayscale value range of 0 ⁇ 255 is divided into 16 areas, and each area is regarded as a color feature interval.
  • 0-15 is a color characteristic interval
  • 16-31 is a color characteristic interval
  • a total of 16 color characteristic intervals b1 to b16 can be obtained.
  • the target gray-scale interval matched by the gray-scale interval b1 is b2
  • the color characteristic range of the gray-scale interval b1 is 0-15
  • the color characteristic value range of the target gray-scale interval b2 is 16-31.
  • the second mapping relationship is: 0 ⁇ 16, 1 ⁇ 17,..., 15 ⁇ 31. Therefore, the gray value of the pixel with the gray value of 0 in the gray level interval b1 in the second histogram is adjusted to 16. Adjust the gray value of the pixel with the gray value of 1 in the gray level interval b1 in the second histogram to 17, and so on.
  • FIG. 5 it is an up-sampled image and a target image corresponding to the up-sampled image obtained by the image processing method provided by the embodiment of the present application. From the comparison of the two images before and after processing in Fig. 5, it can be seen that compared with the up-sampled image, the target image obtained by the image processing method provided by the embodiment of the application does not have obvious aliasing and color edge noise. Quality and display effect due to the original up-sampling graphics, the image details can be better restored.
  • an embodiment of the present application also provides an image processing device 60, which includes a first statistics module 601, a second statistics module 602, a matching module 603, and a processing module. 604.
  • the first statistics module 601 is configured to obtain a first histogram corresponding to the original image according to the number of pixels in the original image whose color feature values fall within each color feature interval.
  • the second statistics module 602 is configured to obtain a second histogram corresponding to the up-sampled image according to the number of pixel points in the up-sampled image of the original image whose color feature values fall within each color feature interval.
  • the matching module 603 is configured to respectively determine the target color feature interval matched by each color feature interval in the second histogram according to the first histogram and the second histogram.
  • the processing module 604 is configured to process the color feature values of the pixels falling in each color feature interval in the up-sampled image according to the target color feature interval matched by each color feature interval in the second histogram to obtain the target image.
  • the matching module 603 is specifically configured to: for any color feature interval in the second histogram, determine the ratio of the proportion value corresponding to each color feature interval in the first histogram and the proportion value corresponding to any color feature interval Determine the target color feature interval matched by any color feature interval from the difference value that meets the specified conditions in the corresponding color feature interval in the first histogram. Among them, each color feature interval in the first histogram is determined.
  • the proportion corresponding to each color feature interval is: the ratio of the number of pixels in the original image whose color feature value falls within the corresponding color feature interval to the total number of pixels contained in the original image, and each color feature in the second histogram
  • the proportion corresponding to the interval is: the ratio of the number of pixels in the up-sampled image whose color feature values fall within the corresponding color feature interval to the total number of pixels contained in the up-sampled image.
  • the matching module 603 is specifically configured to: if there are at least two differences that satisfy the specified condition, determine the color feature interval corresponding to each of the at least two differences in the first histogram, and determine from The color feature interval with the smallest level difference from any one of the color feature intervals is selected from the at least two color feature intervals, and it is determined as the target color feature interval matched by any one of the color feature intervals.
  • the matching module 603 is specifically configured to:
  • the first histogram determine the first mean value and the first variance corresponding to the color feature interval of the pixels in the original image
  • the second histogram determine the second mean value and the second variance corresponding to the color feature interval of the pixels in the up-sampled image
  • the target color feature interval matched by each color feature interval in the second histogram is determined.
  • processing module 604 is specifically configured to:
  • each color feature interval corresponds to a color feature value
  • each color feature interval corresponds to a color feature value range
  • the second mapping relationship between the two is to adjust the color feature value of the pixel points in each color feature interval in the up-sampled image to the color feature value determined according to the second mapping relationship.
  • the color characteristic value includes at least one of the following: the gray value of the pixel, the brightness value of the pixel, and the color value of the pixel.
  • the image processing device provided in the embodiment of the application adopts the same inventive concept as the above-mentioned image processing method, and can achieve the same beneficial effects, which will not be repeated here.
  • an embodiment of the present application also provides an electronic device.
  • the electronic device may specifically be a terminal device or a server in FIG. 1A and FIG. 1B.
  • the electronic device 70 may include a processor 701 and a memory 702.
  • the memory 702 stores computer-readable instructions, and when the computer-readable instructions are executed by the processor 701, the processor 701 executes the steps of the foregoing image processing method.
  • the steps of the image processing method may be the steps in the image processing method of each of the foregoing embodiments.
  • the processor 701 may be a general-purpose processor, such as a central processing unit (CPU), a digital signal processor (Digital Signal Processor, DSP), an application specific integrated circuit (ASIC), and a field programmable gate array (Field Programmable Gate).
  • Array, FPGA or other programmable logic devices, discrete gates or transistor logic devices, and discrete hardware components can implement or execute the methods, steps, and logic block diagrams disclosed in the embodiments of the present application.
  • the general-purpose processor may be a microprocessor or any conventional processor or the like. The steps of the method disclosed in combination with the embodiments of the present application may be directly embodied as being executed and completed by a hardware processor, or executed and completed by a combination of hardware and software modules in the processor.
  • the memory 702 as a non-volatile computer-readable storage medium, can be used to store non-volatile software programs, non-volatile computer-executable programs, and modules.
  • the memory may include at least one type of storage medium, such as flash memory, hard disk, multimedia card, card-type memory, random access memory (Random Access Memory, RAM), static random access memory (Static Random Access Memory, SRAM), Programmable Read Only Memory (PROM), Read Only Memory (ROM), Electrically Erasable Programmable Read-Only Memory (EEPROM), magnetic memory, disk, CD and so on.
  • the memory is any other medium that can be used to carry or store desired program codes in the form of instructions or data structures and that can be accessed by a computer, but is not limited to this.
  • the memory 702 in the embodiment of the present application may also be a circuit or any other device capable of realizing a storage function for storing program instructions and/or data.
  • the embodiment of the present application provides a computer-readable storage medium for storing computer-readable instructions used by the above-mentioned electronic device.
  • the processor executes the steps of the above-mentioned image processing method.
  • a computer program product or computer readable instruction includes a computer readable instruction, and the computer readable instruction is stored in a computer readable storage medium.
  • the processor of the computer device reads the computer-readable instruction from the computer-readable storage medium, and the processor executes the computer-readable instruction, so that the computer device executes the steps in the foregoing method embodiments.
  • the above-mentioned computer-readable storage medium may be any available medium or data storage device that can be accessed by a computer, including but not limited to magnetic storage (such as floppy disk, hard disk, magnetic tape, magneto-optical disk (MO), etc.), optical storage (such as CD, DVD, etc.) , BD, HVD, etc.), and semiconductor memory (such as ROM, EPROM, EEPROM, non-volatile memory (NAND FLASH), solid state drive (SSD)), etc.
  • magnetic storage such as floppy disk, hard disk, magnetic tape, magneto-optical disk (MO), etc.
  • optical storage such as CD, DVD, etc.
  • BD Blu-ray Disc
  • HVD solid state drive
  • semiconductor memory such as ROM, EPROM, EEPROM, non-volatile memory (NAND FLASH), solid state drive (SSD)

Landscapes

  • Engineering & Computer Science (AREA)
  • Theoretical Computer Science (AREA)
  • Physics & Mathematics (AREA)
  • General Physics & Mathematics (AREA)
  • Computer Vision & Pattern Recognition (AREA)
  • Multimedia (AREA)
  • Data Mining & Analysis (AREA)
  • Signal Processing (AREA)
  • Life Sciences & Earth Sciences (AREA)
  • Artificial Intelligence (AREA)
  • Bioinformatics & Cheminformatics (AREA)
  • Bioinformatics & Computational Biology (AREA)
  • Evolutionary Biology (AREA)
  • Evolutionary Computation (AREA)
  • General Engineering & Computer Science (AREA)
  • Image Processing (AREA)
  • Image Analysis (AREA)

Abstract

一种图像处理方法，包括：根据原图像中颜色特征值落入各个颜色特征区间内的像素点的数量，获得原图像对应的第一直方图；根据原图像的上采样图像中颜色特征值落入各个颜色特征区间内的像素点的数量，获得上采样图像对应的第二直方图；根据第一直方图和第二直方图，分别确定出第二直方图中各个颜色特征区间所匹配的目标颜色特征区间；根据第二直方图中各个颜色特征区间所匹配的目标颜色特征区间，对上采样图像中落入各个颜色特征区间的像素点的颜色特征值进行处理，获得目标图像。

Description

图像处理方法、装置、电子设备及存储介质
本申请要求于2020年01月14日提交中国专利局，申请号为202010038997.5，申请名称为“图像处理方法、装置、电子设备及存储介质”的中国专利申请的优先权，其全部内容通过引用结合在本申请中。
技术领域
本申请涉及计算机技术领域，尤其涉及一种图像处理方法、装置、电子设备及存储介质。
背景技术
在图像处理领域中，常常需要获得高分辨率的图像或者还原压缩图像的分辨率，即对图像进行上采样处理。现有的上采样算法主要有插值算法或基于深度学习的算法等，其基本原理都是根据邻近像素点的颜色特征值计算插入的像素点的颜色特征值，由于这些算法只依据局部邻近像素点的颜色信息来计算插入的像素点的颜色信息，导致上采样后的图像会出现明显的锯齿以及色彩边缘噪声等，降低了上采样图像的质量和显示效果。
发明内容
一种图像处理方法，包括：
根据原图像中颜色特征值落入各个颜色特征区间内的像素点的数量，获得所述原图像对应的第一直方图；
根据所述原图像的上采样图像中颜色特征值落入所述各个颜色特征区间内的像素点的数量，获得所述上采样图像对应的第二直方图；
根据所述第一直方图和所述第二直方图，分别确定出所述第二直方图中各个颜色特征区间所匹配的目标颜色特征区间；及
根据所述第二直方图中各个颜色特征区间所匹配的目标颜色特征区间，对所述上采样图像中落入所述各个颜色特征区间的像素点的颜色特征值进行 处理，获得目标图像。
一种图像处理装置，包括：
第一统计模块，用于根据原图像中颜色特征值落入各个颜色特征区间内的像素点的数量，获得所述原图像对应的第一直方图；
第二统计模块，用于根据所述原图像的上采样图像中颜色特征值落入所述各个颜色特征区间内的像素点的数量，获得所述上采样图像对应的第二直方图；
匹配模块，用于根据所述第一直方图和所述第二直方图，分别确定出所述第二直方图中各个颜色特征区间所匹配的目标颜色特征区间；及
处理模块，用于根据所述第二直方图中各个颜色特征区间所匹配的目标颜色特征区间，对所述上采样图像中落入所述各个颜色特征区间的像素点的颜色特征值进行处理，获得目标图像。
一种电子设备，包括存储器和一个或多个处理器，存储器中储存有计算机可读指令，计算机可读指令被处理器执行时，使得一个或多个处理器执行上述图像处理方法的步骤。
一个或多个存储有计算机可读指令的非易失性计算机可读存储介质，计算机可读指令被一个或多个处理器执行时，使得一个或多个处理器执行上述图像处理方法的步骤。
附图说明
为了更清楚地说明本申请实施例的技术方案，下面将对本申请实施例中所需要使用的附图作简单地介绍，显而易见地，下面所介绍的附图仅仅是本申请的一些实施例，对于本领域普通技术人员来讲，在不付出创造性劳动的前提下，还可以根据这些附图获得其他的附图。
图1A为本申请一个实施例提供的图像处理方法的应用场景示意图；
图1B为本申请另一个实施例提供的图像处理方法的应用场景示意图；
图1C为本申请一个实施例提供的图像处理方法的应用于神经网络的示意图；
图1D为本申请另一个实施例提供的图像处理方法的应用于神经网络的示意图；
图2为本申请一实施例提供的图像处理方法的流程示意图；
图3为本申请一实施例提供的图像对应的灰度值以及统计灰度值得到的直方图；
图4为本申请一实施例提供的确定出第二直方图中各个颜色特征区间所匹配的目标颜色特征区间的流程示意图；
图5为本申请一实施例提供的上采样图像和目标图像的对应图；
图6为本申请一实施例提供的图像处理装置的结构示意图；
图7为本申请一实施例提供的电子设备的结构示意图。
具体实施方式
为使本申请实施例的目的、技术方案和优点更加清楚，下面将结合本申请实施例中的附图，对本申请实施例中的技术方案进行清楚、完整地描述。
为了方便理解，下面对本申请实施例中涉及的名词进行解释：
直方图(Histogram)，又称质量分布图，是一种统计报告图，由一系列高度不等的纵向条纹或线段表示数据分布的情况，一般用横轴表示数据类型，纵轴表示分布情况。在本申请中，预先将颜色特征值划分成多个颜色特征区间，即直方图的横轴为颜色特征区间，统计一个图像中的像素点的颜色特征值落入各个颜色特征区间中的数量，得到表征该图像的颜色特征分布情况的直方图。
RGB：是工业界的一种颜色标准，是通过对红(R)、绿(G)、蓝(B)三个颜色通道的变化以及它们相互之间的叠加来得到各式各样的颜色的，RGB即是代表红、绿、蓝三个通道的颜色，这个标准几乎包括了人类视力所能感知的 所有颜色，是运用最广的颜色系统之一。
YUV：是一种颜色编码方法，常使用在各个视频处理组件中。YUV在对照片或视频编码时，考虑到人类的感知能力，允许降低色度的带宽。YUV是编译true-color颜色空间(color space)的种类，Y'UV、YUV、YCbCr、YPbPr等都可以称为YUV，彼此有重叠。“Y”表示明亮度(Luminance或Luma)，也就是灰阶值，“U”和“V”表示的则是色度(Chrominance或Chroma)，作用是描述影像色彩及饱和度，用于指定像素的颜色。
CIELab：CIELab是CIE的一个颜色系统，表色体系，基于CIELab的意思是基于这个颜色系统之上，基本是用于确定某个颜色的数值信息。Lab模式是由国际照明委员会(CIE)于1976年公布的一种色彩模式。是CIE组织确定的一个理论上包括了人眼可见的所有色彩的色彩模式。Lab模式弥补了RGB与CMYK两种彩色模式的不足，是Photoshop用来从一种色彩模式向另一种色彩模式转换时使用的一种内部色彩模式。Lab模式也是由三个通道组成，第一个通道是明度，即“L”。a通道的颜色是从红色到深绿；b通道则是从蓝色到黄色。在表达色彩范围上，最全的是Lab模式，其次是RGB模式，最窄的是CMYK模式。也就是说Lab模式所定义的色彩最多，且与光线及设备无关，并且处理速度与RGB模式同样快，比CMYK模式快数倍。
颜色特征区间：是对颜色特征对应的颜色特征值的范围进行划分后得到的多个区间范围。本申请中，颜色特征可以是像素点的灰度、亮度、颜色等等表示像素点色彩特征的数据。例如，例如，颜色特征为像素点的灰度，灰度值的范围为0～255，可以将0～255的灰度值范围划分成多个区间范围，每个区间范围作为一个颜色特征区间，如0～15为一个颜色特征区间，16～31为一个颜色特征区间，以此类推一共可得到16个连续的颜色特征区间b1～b16，为方便描述，可将颜色特征区间bi的等级记为i。
下采样(subsampled)：也可以称为降采样(downsampled)，其目的是降低图像的分辨率，即缩减图像的像素点数量，使得图像符合显示区域的大小 或生成原图像对应的缩略图。
上采样(upsampling)：也可以为图像超分辨率(Image Super Resolution)，其目的是提高图像的分别率，即增加图像的像素点数量，使得图像可以适应高分辨率的应用场合或者恢复原图像丢失的细节。实际应用中，可通过图像插值(interpolating)或基于深度学习的算法对图像进行上采样处理。常见的插值算法有：最近邻插值、双线性插值、双三次插值算法、均值插值、中值插值等方法。
通常在需要获取到高分辨率的图像时，可对原图像进行上采样，以还原或重建原图像中更多的细节，进而在更高分辨率的显示设备进行显示。或者，在利用神经网络对图像进行处理时，为了保证输入图像的分别率适合神经网络的输入大小，通常会先对输入图像进行下采样后再输入神经网络，然后对神经网络的输出图像进行上采样，以将输出图像的还原至原始的分辨率；在神经网络中也可设置至少一个下采样网络层，通过在处理过程中降图像的分辨率来提高处理效率，并通过设置至少一个上采样网络层，还原下采样图像的分辨率。
附图中的任何元素数量均用于示例而非限制，以及任何命名都仅用于区分，而不具有任何限制含义。
在具体实践过程中，相关的上采样算法主要有图像插值算法或基于深度学习的算法等，任何一种插值算法的基本原理都是根据邻近像素点的颜色特征值计算插入的像素点的颜色特征值，而基于深度学习的算法的基本原理也是通过神经网络学习邻近像素点之间的关系得到插入的像素点，由于这些算法只依据局部邻近像素点的颜色信息来计算插入的像素点的颜色信息，导致上采样后的图像会出现明显的锯齿以及色彩边缘噪声等，且图像边缘处会出现缺上、下、左、右其中一个或二个方向的邻近像素集，目前的算法只能用可用邻近像素点进行填充，导致图像边缘处的像素点效果与预期相差大，这些都降低了上采样图像的质量和显示效果。
为此，本申请提出了一种图像处理方法，包括：根据原图像中颜色特征值落入各个颜色特征区间内的像素点的数量，获得原图像对应的第一直方图；根据原图像的上采样图像中颜色特征值落入各个颜色特征区间内的像素点的数量，获得上采样图像对应的第二直方图；根据第一直方图和第二直方图，分别确定出第二直方图中各个颜色特征区间所匹配的目标颜色特征区间；根据各个颜色特征区间所匹配的目标颜色特征区间，对上采样图像中落入各个颜色特征区间的像素点的颜色特征值进行处理，获得目标图像。上述图像处理方法中，第一直方图表征原图像中颜色特征分布情况的直方图，第二直方图表征上采样图像的颜色特征分布情况的直方图，根据第一直方图和第二直方图揭示的颜色特征分布情况，确定出第二直方图中各个颜色特征区间所匹配的目标颜色特征区间，然后根据所匹配的目标颜色特征区间，对上采样图像中的像素点的颜色特征值逐一进行调整，得到上采样图像对应的目标图像，使得上采样图像对应的目标图像的颜色特征分布情况尽可能的与原图像的颜色特征分布情况一致，该方法能够明显削弱上采样图像中出现的锯齿、色彩边缘噪声等，提高上采样图像的质量和显示效果，更好地还原图像细节，且处理方式简单高效，尤其适用于对处理效率要求高的应用场合，如视频实时传输。
在介绍完本申请实施例的设计思想之后，下面对本申请实施例的技术方案能够适用的应用场景做一些简单介绍，需要说明的是，以下介绍的应用场景仅用于说明本申请实施例而非限定。在具体实施时，可以根据实际需要灵活地应用本申请实施例提供的技术方案。
参考图1A，其为本申请实施例提供的图像处理方法的应用场景示意图。该应用场景包括多个终端设备，包括终端设备101-1、终端设备101-2、……终端设备101-n)、后台服务器102。其中，各个终端设备和后台服务器102之间通过无线或有线网络连接，终端设备包括但不限于桌面计算机、智能手机、移动电脑、平板电脑、媒体播放器、智能可穿戴设备、电视、监控摄像 头等。后台服务器可以是一台服务器、若干台服务器组成的服务器集群或云计算中心。终端设备可将图像或视频上传到后台服务器102，或者从后台服务器102中获取到图像或视频。在图像或视频采集或传输过程中，为了有效利用有限的带宽，可以传输低分辨率的图像或视频，然后在接收端将低分辨率的图像或视频转换成高分辨率图像或视频，终端设备或后台服务器102均可以在对图像或视频进行上采样处理时结合本申请提供的图像处理方法对上采样后的图像进行增强处理。具体的应用场景不限于视频直播、视频传输、图像传输等。
当然，也可以直接在两台终端设备之间传输图像或视频，例如监控摄像头将采集到的视频下采样处理后传输给监控显示器，监控显示器对下采样的视频进行上采样处理，获得高分辨率的视频进行显示。
参考图1B，其为本申请实施例提供的图像处理方法的应用场景示意图。其中，终端设备111中安装有图像处理应用，该图像应用中植入了本申请提供的图像处理方法对应的软件功能模块。在用户使用终端设备111内的图像处理应用获得原图像对应的高分辨率图像(即上采样图像)时，可通过本申请提供的图像处理方法，基于原图像的直方图和上采样图像的直方图，对上采样图像进行处理，获得高分辨率的图像。例如，被压缩后的图像或视频、被污损的照片、分辨率较低的监控视频和照片等，可以通过上述图像处理应用对图像进行恢复和重建。终端设备111包括但不限于桌面计算机、移动电脑、平板电脑、智能手机等可运行图像处理应用的设备。
本申请提供的图像处理方法还可以植入在图像查看软件或图像编辑软件中。在用户通过图像查看软件或图像编辑软件查看图像过程中，如果用户需要专注于图像的某些细节时，可对图像进行缩放变换，此时，图像查看软件或图像编辑软件可基于原图像的直方图和上采样图像的直方图，对上采样图像进行处理，获得高分辨率的图像。
本申请提供的图像处理方法，可广泛应用于雷达图像、卫星遥感图像、 天文观测图像、地质勘探数据图像、生物医学切片、显微图像等特殊图像，日常人物景物图像以及视频的处理。
本申请提供的图像处理方法，还可以应用于神经网络中。例如，参考图1C,在利用神经网络对图像进行处理时，为了保证输入图像的分辨率适合神经网络的输入大小，通常会先对输入图像进行下采样后再输入神经网络，然后对神经网络的输出图像进行上采样，以将输出图像的还原至输入图像的分辨率，可利用本申请提供的图像处理方法对上采样图像进行增强处理，得到目标图像，以提高目标图像的质量和显示效果。又如，参考图1D，神经网络中还可以包括至少一次下采样处理，通过在处理过程中降低图像的分辨率来提高处理效率，并通过至少一次上采样处理，将还原图像的分辨率，此时，可在上采样网络层后增加一个增强处理层，以利用本申请提供的图像处理方法对上采样处理后的图像进行处理，以提高上采样图像的质量，将处理后的图像输入下一个网络层进行处理。
当然，本申请实施例提供的方法并不限用于上述应用场景中，还可以用于其它可能的应用场景，本申请实施例并不进行限制。对于上述应用场景的各个设备所能实现的功能将在后续的方法实施例中一并进行描述，在此先不过多赘述。
为进一步说明本申请实施例提供的技术方案，下面结合附图以及具体实施方式对此进行详细的说明。虽然本申请实施例提供了如下述实施例或附图所示的方法操作步骤，但基于常规或者无需创造性的劳动在所述方法中可以包括更多或者更少的操作步骤。在逻辑上不存在必要因果关系的步骤中，这些步骤的执行顺序不限于本申请实施例提供的执行顺序。
下面结合上述应用场景，对本申请实施例提供的技术方案进行说明。
参考图2，本申请实施例提供的一种图像处理方法，可应用于上述应用场景中的终端设备或后台服务器，具体包括以下步骤：
S201、根据原图像中颜色特征值落入各个颜色特征区间内的像素点的数 量，获得原图像对应的第一直方图。
本申请实施例中，原图像可以是指上采样处理前的图像，原图像可以是一幅单独的图像，也可以视频中一帧一帧的图像。
本申请实施例中的颜色特征是指表征像素点色彩特性的特征，例如可以是像素点的灰度、亮度、颜色等。相应地，颜色特征值包括以下至少一种：像素点的灰度值、像素点的亮度值、像素点的颜色值等。
具体实施时，颜色值可根据图像采用的颜色系统确定。例如，当图像为灰度图像时，图像中每个像素点的颜色特征可包括灰度值，此时针对一个图像可得到灰度值对应的直方图。当图像采用RGB描述色彩时，图像中每个像素点的颜色特征包括R(红色)、G(绿色)、B(蓝色)三个颜色特征，可对这三个颜色特征分别进行统计，即针对一个图像可得到针对R(红色)、G(绿色)、B(蓝色)三个颜色特征分别对应的三个直方图。对于彩色图像，可通过如下公式得到彩色图像中每个像素点的灰度值：gray＝R×0.299+G×0.587+B×0.114。
具体实施时，颜色特征区间可由本领域技术人员根据具体应用场景以及颜色特征对应的特征值的范围预先设定，一个颜色特征区间可以对应一个颜色特征值或者一段颜色特征值范围。例如，颜色特征为像素点的灰度，灰度值的范围为0～255，可将每一个灰度值作为一个颜色特征区间，即一共可得到256个灰度区间；也可以将0～255的灰度值范围划分成多个区域，每个区域作为一个颜色特征区间，如0～15为一个颜色特征区间，16～31为一个颜色特征区间，以此类推一共可得到16个颜色特征区间b1～b16，参考图3，其示出了一个图像中每个像素点的灰度值，针对该图像，统计灰度值落入上述16个颜色特征区间内的像素点数量，得到该图像的灰度值对应的直方图。
S202、根据原图像的上采样图像中颜色特征值落入各个颜色特征区间内的像素点的数量，获得上采样图像对应的第二直方图。
本申请实施例中的上采样图像为对原图像进行上采样处理后得到的图像，本申请对具体的上采样处理方法不作限定。
需要说明的是，第一直方图和第二直方图中的颜色特征区间相同。
S203、根据第一直方图和第二直方图，分别确定出第二直方图中各个颜色特征区间所匹配的目标颜色特征区间。
S204、根据第二直方图中各个颜色特征区间所匹配的目标颜色特征区间，对上采样图像中落入各个颜色特征区间的像素点的颜色特征值进行处理，获得目标图像。
本申请实施例中，第一直方图表征原图像中颜色特征分布情况的直方图，第二直方图表征上采样图像的颜色特征分布情况的直方图。
具体实施时，根据第一直方图和第二直方图揭示的颜色特征分布情况，可确定出第二直方图中每个颜色特征区间和第一直方图中的颜色特征区间之间的匹配关系，该匹配关系揭示了第一直方图和第二直方图中颜色特征分布情况最接近的颜色特征区间之间的对应关系，基于匹配关系确定出第二直方图中各个颜色特征区间所匹配的目标颜色特征区间。然后，根据第二直方图中各个颜色特征区间所匹配的目标颜色特征区间，对上采样图像中的像素点的颜色特征值逐一进行调整，得到上采样图像对应的目标图像，使得上采样图像对应的目标图像的颜色特征分布情况尽可能的与原图像的颜色特征分布情况一致。
具体实施时，若第二直方图中某一颜色特征区间与其所匹配目标颜色特征区间相同，则不需要对上采样图像中落入该颜色特征区间的像素点的颜色特征值进行调整。
针对一幅原图像的上采样图像，可同时对多个颜色特征进行调整，此时针对这多个颜色特征中的每个颜色特征，获得原图像的针对每个颜色特征的第一直方图，以及上采样图像的针对每个颜色特征的第二直方图，然后，分别基于各个颜色特征对应的第一直方图和第二直方图，对上采样图像中的各 个颜色特征的颜色特征值进行调整。
例如，需要对上采样图像中的亮度和灰度进行调整，则获得原图像的亮度对应的第一直方图VL1、原图像的灰度对应的第一直方图VG1、上采样图像的亮度对应的第二直方图VL2，上采样图像的灰度对应的第二直方图VG2。然后，基于第一直方图VL1和第二直方图VL2，分别确定出第二直方图中各个亮度区间所匹配的目标亮度区间，根据各个亮度区间所匹配的目标亮度区间，将上采样图像P1中落入各个亮度区间的像素点的亮度值调整为各个亮度区间所匹配的目标亮度区间对应的亮度值，从而获得图像P2。接着，基于第一直方图VG1和第二直方图VG2，分别确定出第二直方图中各个灰度区间所匹配的目标灰度区间，根据各个灰度区间所匹配的目标灰度区间，将图像P2中落入各个灰度区间的像素点的灰度值调整为各个灰度区间所匹配的目标灰度区间对应的灰度值，从而获得目标图像P3。
例如，当图像采样YUV颜色编码时，可获得原图像的颜色特征Y、U、V分别对应的第一直方图，以及上采样图像的颜色特征Y、U、V分别对应的第二直方图。基于颜色特征Y对应的第一直方图和第二直方图，分别确定出颜色特征Y对应的第二直方图中各个颜色特征区间所匹配的目标颜色特征区间，根据各个颜色特征区间所匹配的目标颜色特征区间，对上采样图像P1中落入颜色特征Y对应的第二直方图的各个颜色特征区间中的像素点的颜色特征Y的数值进行处理，获得图像P2。然后，基于颜色特征U对应的第一直方图和第二直方图，分别确定出颜色特征U对应的第二直方图中各个颜色特征区间所匹配的目标颜色特征区间，根据各个颜色特征区间所匹配的目标颜色特征区间，对图像P2中落入颜色特征U对应的第二直方图的各个颜色特征区间中的像素点的颜色特征U的数值进行处理，获得图像P3。最后，基于颜色特征V对应的第一直方图和第二直方图，分别确定出颜色特征V对应的第二直方图中各个颜色特征区间所匹配的目标颜色特征区间，根据各个颜色特征区间所匹配的目标颜色特征区间，对图像P3中落入颜色特征V对应的 第二直方图的各个颜色特征区间中的像素点的颜色特征V的数值进行处理，获得目标图像。
本申请实施例提供的图像处理方法，能够明显削弱上采样图像中出现的锯齿、色彩边缘噪声等，提高上采样图像的质量和显示效果，更好地还原图像细节，且处理方式简单高效，尤其适用于对处理效率要求高的应用场合，如视频实时传输。
具体实施时，可通过多种方式确定出第二直方图中每个颜色特征区间和第一直方图中的颜色特征区间之间的匹配关系，从而确定第二直方图中各个颜色特征区间所匹配的目标颜色特征区间。
在一种可能的实施方式中，可根据第一直方图和第二直方图分别统计的各个颜色特征区间对应的占比值，确定第二直方图中各个颜色特征区间所匹配的目标颜色特征区间。其中，第一直方图中的每个颜色特征区间对应的占比值为：原图像中颜色特征值落入对应的颜色特征区间内的像素点的数量与原图像包含的像素点总数的比值，第二直方图中的每个颜色特征区间对应的占比值为：上采样图像中颜色特征值落入对应的颜色特征区间内的像素点的数量与上采样图像包含的像素点总数的比值。
基于此，步骤S203具体包括：针对第二直方图中的任一颜色特征区间，确定第一直方图中各个颜色特征区间对应的占比值和任一颜色特征区间对应的占比值之间的差值，从满足指定条件的差值在第一直方图中对应的颜色特征区间中，确定出任一颜色特征区间所匹配的目标颜色特征区间。
具体实施时，指定条件可以是第二直方图中的任一颜色特征区间对应的所有差值中的最小值。例如，第一直方图和第二直方图均包含16个颜色特征区间b1～b16，以第二直方图中的颜色特征区间b1为例，分别计算第二直方图中颜色特征区间b1的占比值和第一直方图中每个颜色特征区间对应的占比值之间的差值，一共可得到16个差值，从这16个差值中选出最小的差值，将该差值在第一直方图中对应的颜色特征区间确定为第二直方图中颜色特征 区间b1所匹配的目标颜色特征区间，例如，最小的差值为第二直方图中的颜色特征区间b1的占比值和第一直方图中颜色特征区间b3的占比值之间的差值，则第二直方图中颜色特征区间b1所匹配的目标颜色特征区间为b3。
具体实施时，也可以按从小到大的顺序，对第二直方图中的任一颜色特征区间对应的所有差值进行排序，此时指定条件可以是排序靠前的N个差值，其中N大于等于1，且小于颜色特征区间的总数，N的具体取值可由本领域技术人员根据实际应用场景确定，本申请实施例不作限定。例如，第一直方图和第二直方图均包含16个颜色特征区间b1～b16，且N＝3，以第二直方图中的颜色特征区间b1为例，分别计算第二直方图中颜色特征区间b1的占比值和第一直方图中每个颜色特征区间对应的占比值之间的差值，一共可得到16个差值，按从小到大的顺序对这16个差值进行排序，选取排序靠前的3个差值，假设排序靠前的3个差值在第一直方图中对应的颜色特征区间为b1、b3、b10，则从b1、b3、b10中确定出一个颜色特征区间作为第二直方图中颜色特征区间b1所匹配的目标颜色特征区间。
具体实施时，指定条件还可以是差值小于差值阈值，即将小于差值阈值的差值在第一直方图中对应的颜色特征区间，确定为该差值在第二直方图中对应的颜色特征区间所匹配的目标颜色特征区间。以第二直方图中的颜色特征区间b1为例，第二直方图中的颜色特征区间b1的占比值和第一直方图中颜色特征区间b4的占比值之间的差值小于差值阈值，第二直方图中的颜色特征区间b1的占比值和第一直方图中颜色特征区间b2的占比值之间的差值小于差值阈值，则从颜色特征区间b2和b4中确定出一个颜色特征区间作为第二直方图中颜色特征区间b1所匹配的目标颜色特征区间。
上述实施方式提供的图像处理方法，可根据第一直方图和第二直方图中各个颜色特征区间之间的占比值的差值，简单高效地确定第二直方图中各个颜色特征区间所匹配的目标颜色特征区间，能够明显削弱上采样图像中出现的锯齿、色彩边缘噪声等，提高上采样图像的质量和显示效果，从而更好地 还原图像细节。
进一步地，针对第二直方图中的任一颜色特征区间，若仅存在一个满足指定条件的差值，则将该满足指定条件的差值在第一直方图中对应的颜色特征区间确定为该任一颜色特征区间所匹配的目标颜色特征区间；若存在至少两个差值满足指定条件，则确定至少两个差值中每个差值在第一直方图中对应的颜色特征区间，从确定出的至少两个颜色特征区间中选出与该任一颜色特征区间的等级差最小的颜色特征区间，确定为该任一颜色特征区间所匹配的目标颜色特征区间。
本申请实施例中，两个颜色特征区间之间的等级差越小，表明这两个颜色特征区间所代表的颜色特征值越相近。以灰度值为例，0～15为一个颜色特征区间b1对应的灰度值范围为0～15，颜色特征区间b2对应的灰度值范围为16～31，颜色特征区间b16对应的灰度值范围为240～255，显然，颜色特征区间b1和颜色特征区间b2之间的灰度值更接近。由于同一图像中可能存在多个颜色特征区间的占比值相同的情况，因此针对第二直方图中的一个颜色特征区间，在第一直方图中可能存在多个颜色特征区间的占比值与该第二直方图中的该颜色特征区间的占比值相同或相近，而相近的颜色特征区间才是期望的调整后的目标颜色特征区间。
以第二直方图中的颜色特征区间b1为例，分别计算第二直方图中颜色特征区间b1的占比值和第一直方图中每个颜色特征区间对应的占比值之间的差值，一共可得到16个差值：d 1,1、d 1,2、…d 1,i、…d 1,16，其中，d 1,i表示第二直方图中的颜色特征区间b1的占比值和第二直方图中的颜色特征区间bi的占比值之间的差值。如果上述16个差值中的d 1,2和d 1,5满足指定条件，则第二直方图中的颜色特征区间b1在第一直方图中所匹配的颜色特征区间分别为b2和b5，此时，颜色特征区间b2与颜色特征区间b1的等级差小于颜色特征区间b5与颜色特征区间b1的等级差，即颜色特征区间b2所对应的颜色特征值更接近颜色特征区间b1所对应的颜色特征值，则第二直方图中颜色特 征区间b1所匹配的目标颜色特征区间为b2。
进一步地，基于上述直方图匹配的发明构思，为了提高匹配准确度，可基于第一直方图和第二直方图统计得到的颜色特征值的分布情况，通过一些算法确定出第一直方图和第二直方图中的颜色特征区间之间的映射关系，基于映射关系可确定出第二直方图中各个颜色特征区间所匹配的目标颜色特征区间。
例如，可通过如下公式确定出第一直方图和第二直方图中的颜色特征区间之间的映射关系：
Figure PCTCN2020122203-appb-000001
其中，MIN为求最小值的函数，Index为获得函数MIN输出的最小值对应的颜色特征区间bj的函数，i表示第二直方图V2中的第i个颜色特征区间bi，V2[i]表示第二直方图V2中第i个颜色特征区间bi的占比值，V1[j]表示第一直方图V1中第j个颜色特征区间bj的占比值，1≤j≤n，n为颜色特征区间的总数。其中，a＝MSE/618，其中均方误差
Figure PCTCN2020122203-appb-000002
参数a可表征第一直方图和第二直方图之间的颜色特征区间的占比值的整体偏移情况，参数c＝|j-i|，表征第一直方图中的颜色特征区间j和第二直方图中的颜色特征区间i之间的等级差，增加参数c可提高与第二直方图中颜色特征区间i接近的颜色特征区间被命中的概率。F(i)即为第二直方图V2中的第i个颜色特征区间bi所匹配的目标颜色特征区间对应的索引号，例如F(i)＝k，第二直方图V2中的第i个颜色特征区间bi所匹配的目标颜色特征区间bk。
实际应用中，映射关系F(i)不限于上述列举的公式，参数a和c也不限于上述列举的方式。例如，映射关系还可以是：
Figure PCTCN2020122203-appb-000003
或者，还可以是：
Figure PCTCN2020122203-appb-000004
其中，MAX为求最大值的函数，保证F(i)≥1。
上述实施方式提供的图像处理方法，在第一直方图和第二直方图中各个颜色特征区间之间的占比值的差值的基础上，结合颜色特征区间之间的等级差，确定出第二直方图中各个颜色特征区间所匹配的目标颜色特征区间，进一步地提高了匹配准确度。
在另一种可能的实施方式中，可根据原图像的第一直方图和上采样图像的第二直方图，计算第二直方图中各个颜色特征区间所匹配的目标颜色特征区间。其中，第一直方图统计的是原图像中颜色特征值落入各个颜色特征区间内的像素点的数量，第二直方图统计的是上采样图像中颜色特征值落入各个颜色特征区间内的像素点的数量。
基于此，参考图4，步骤S203可包括如下步骤：
S401、根据第一直方图确定原图像中像素点的颜色特征区间对应的第一均值和第一方差，以及根据第二直方图确定上采样图像中像素点的颜色特征区间对应的第二均值和第二方差。
具体实施时，第一均值
Figure PCTCN2020122203-appb-000005
第一方差
Figure PCTCN2020122203-appb-000006
第二均值
Figure PCTCN2020122203-appb-000007
第一方差
Figure PCTCN2020122203-appb-000008
其中，n为颜色特征区间的总数，V1[i]表示第一直方图V1中第i个颜色特征区间bi的占比值，V2[i]表示第二直方图V2中第i个颜色特征区间bi的占比值。
S402、根据第一均值、第一方差、第二均值和第二方差，确定第一直方图中的颜色特征区间和第二直方图中的颜色特征区间之间的第一映射关系。
具体实施时，第一映射关系可以是：
F(i)＝round(α(i-E2)+E1)
其中，α＝D2/D1，round为四舍五入求整数的函数，V2(i)为第二直方图中落入颜色特征区间bi内的像素点的数量。F(i)即为第二直方图V2中的第i个颜色特征区间bi所匹配的目标颜色特征区间对应的索引号，例如F(1)＝2，则第二直方图V2中颜色特征区间b1所匹配的目标颜色特征区间为b2。
具体实施时，第一映射关系也可以是：
F(i)＝MID(round(α(i-E2)+E1)，1，n)
其中，α＝D2/D1，MID为求round(α(V2[i]-E2)+E1)，1和n这三个数的中间值的函数，保证1≤F(i)≤n，round为四舍五入求整数的函数。
S403、根据第一映射关系，确定第二直方图中各个颜色特征区间所匹配的目标颜色特征区间。
具体实施时，假设F(1)＝2，则第二直方图V2中颜色特征区间b1所匹配的目标颜色特征区间为b2，F(2)＝4，则第二直方图V2中颜色特征区间 b2所匹配的目标颜色特征区间为b4，以此类推得到第二直方图中每个颜色特征区间所匹配的目标颜色特征区间。
上述实施方式提供的图像处理方法，直接根据第一直方图和第二直方图中各个颜色特征区间之间的占比值，计算第一直方图和第二直方图中各个颜色特征区间之间的映射关系，根据映射关系确定出第二直方图中各个颜色特征区间所匹配的目标颜色特征区间，处理方式简单高效。
在上述任一实施方式的基础上，步骤S204具体包括：若每个颜色特征区间对应一个颜色特征值，则分别将上采样图像中落入各个颜色特征区间的像素点的颜色特征值，调整为各个颜色特征区间所匹配的目标颜色特征区间对应的颜色特征值，获得目标图像。
以颜色特征为像素点的灰度为例进行说明，假设灰度值的范围为0～255，将每一个灰度值作为一个颜色特征区间，即一共可得到256个灰度区间b1～b256，即每个灰度区间对应一个灰度值。假设，第二直方图中灰度区间b1所匹配的目标灰度区间为b2，则将上采样图像中落入灰度区间b1的像素点的灰度值0改为目标灰度区间b2对应的灰度值1；第二直方图中灰度区间b2所匹配的目标灰度区间为b2，则无需对上采样图像中落入灰度区间b2的像素点的灰度值进行调整。
在上述任一实施方式的基础上，步骤S204具体包括：若每个颜色特征区间对应一个颜色特征值范围，则分别确定第二直方图中各个颜色特征区间对应的颜色特征值范围和各个颜色特征区间所匹配的目标颜色特征区间对应的颜色特征值范围之间的第二映射关系，将上采样图像中落入各个颜色特征区间的像素点的颜色特征值调整为根据第二映射关系确定的颜色特征值。
以颜色特征为像素点的灰度为例进行说明，假设灰度值的范围为0～255，将0～255的灰度值范围划分成多个16个区域，每个区域作为一个颜色特征区间，如0～15为一个颜色特征区间，16～31为一个颜色特征区间，以此类推一共可得到16个颜色特征区间b1～b16。第二直方图中灰度区间b1所匹配的目 标灰度区间为b2，灰度区间b1的颜色特征范围为0～15，目标灰度区间b2的颜色特征值范围为16～31，两者之间的第二映射关系为：0→16，1→17，……，15→31，因此，将第二直方图中灰度区间b1中灰度值为0的像素点的灰度值调整为16，将第二直方图中灰度区间b1中灰度值为1的像素点的灰度值调整为17，以此类推。
参考图5，其为一幅上采样图像，以及通过本申请实施例提供的图像处理方法获得的该上采样图像对应的目标图像。从图5中处理前后的两幅图像的对比可知，与上采样图像相比，通过本申请实施例提供的图像处理方法获得的目标图像，不会出现明显的锯齿以及色彩边缘噪声，目标图像的质量和显示效果由于原始的上采样图形，能够更好地还原图像细节。
如图6所示，基于与上述图像处理方法相同的发明构思，本申请实施例还提供了一种图像处理装置60，包括第一统计模块601、第二统计模块602、匹配模块603和处理模块604。
第一统计模块601，用于根据原图像中颜色特征值落入各个颜色特征区间内的像素点的数量，获得原图像对应的第一直方图。
第二统计模块602，用于根据原图像的上采样图像中颜色特征值落入各个颜色特征区间内的像素点的数量，获得上采样图像对应的第二直方图。
匹配模块603，用于根据第一直方图和第二直方图，分别确定出第二直方图中各个颜色特征区间所匹配的目标颜色特征区间。
处理模块604，用于根据第二直方图中各个颜色特征区间所匹配的目标颜色特征区间，对上采样图像中落入各个颜色特征区间的像素点的颜色特征值进行处理，获得目标图像。
可选地，匹配模块603具体用于：针对第二直方图中的任一颜色特征区间，确定第一直方图中各个颜色特征区间对应的占比值和任一颜色特征区间对应的占比值之间的差值，从满足指定条件的差值在第一直方图中对应的颜色特征区间中，确定出任一颜色特征区间所匹配的目标颜色特征区间，其中， 第一直方图中的每个颜色特征区间对应的占比值为：原图像中颜色特征值落入对应的颜色特征区间内的像素点的数量与原图像包含的像素点总数的比值，第二直方图中的每个颜色特征区间对应的占比值为：上采样图像中颜色特征值落入对应的颜色特征区间内的像素点的数量与上采样图像包含的像素点总数的比值。
可选地，匹配模块603具体用于：若存在至少两个差值满足指定条件，则确定至少两个差值中每个差值在第一直方图中对应的颜色特征区间，从确定出的至少两个颜色特征区间中选出与任一颜色特征区间的等级差最小的颜色特征区间，确定为任一颜色特征区间所匹配的目标颜色特征区间。
可选地，匹配模块603具体用于：
根据第一直方图，确定原图像中像素点的颜色特征区间对应的第一均值和第一方差；
根据第二直方图，确定上采样图像中像素点的颜色特征区间对应的第二均值和第二方差；
根据第一均值、第一方差、第二均值和第二方差，确定第一直方图中的颜色特征区间和第二直方图中的颜色特征区间之间的第一映射关系；
根据第一映射关系，确定第二直方图中各个颜色特征区间所匹配的目标颜色特征区间。
可选地，处理模块604具体用于：
若每个颜色特征区间对应一个颜色特征值，则分别将上采样图像中落入各个颜色特征区间的像素点的颜色特征值，调整为各个颜色特征区间所匹配的目标颜色特征区间对应的颜色特征值，获得目标图像；
若每个颜色特征区间对应一个颜色特征值范围，则分别确定第二直方图中各个颜色特征区间对应的颜色特征值范围和各个颜色特征区间所匹配的目标颜色特征区间对应的颜色特征值范围之间的第二映射关系，将上采样图像中落入各个颜色特征区间的像素点的颜色特征值调整为根据第二映射关系确 定的颜色特征值。
可选地，颜色特征值包括以下至少一种：像素点的灰度值、像素点的亮度值、像素点的颜色值。
本申请实施例提的图像处理装置与上述图像处理方法采用了相同的发明构思，能够取得相同的有益效果，在此不再赘述。
基于与上述图像处理方法相同的发明构思，本申请实施例还提供了一种电子设备，该电子设备具体可以为图1A、图1B中的终端设备或服务器等。如图7所示，该电子设备70可以包括处理器701和存储器702。存储器702存储有计算机可读指令，计算机可读指令被处理器701执行时，使得处理器701执行上述图像处理方法的步骤。此处图像处理方法的步骤可以是上述各个实施例的图像处理方法中的步骤。
处理器701可以是通用处理器，例如中央处理器(CPU)、数字信号处理器(Digital Signal Processor，DSP)、专用集成电路(Application Specific Integrated Circuit，ASIC)、现场可编程门阵列(Field Programmable Gate Array，FPGA)或者其他可编程逻辑器件、分立门或者晶体管逻辑器件、分立硬件组件，可以实现或者执行本申请实施例中公开的各方法、步骤及逻辑框图。通用处理器可以是微处理器或者任何常规的处理器等。结合本申请实施例所公开的方法的步骤可以直接体现为硬件处理器执行完成，或者用处理器中的硬件及软件模块组合执行完成。
存储器702作为一种非易失性计算机可读存储介质，可用于存储非易失性软件程序、非易失性计算机可执行程序以及模块。存储器可以包括至少一种类型的存储介质，例如可以包括闪存、硬盘、多媒体卡、卡型存储器、随机访问存储器(Random Access Memory，RAM)、静态随机访问存储器(Static Random Access Memory，SRAM)、可编程只读存储器(Programmable Read Only Memory，PROM)、只读存储器(Read Only Memory，ROM)、带电可 擦除可编程只读存储器(Electrically Erasable Programmable Read-Only Memory，EEPROM)、磁性存储器、磁盘、光盘等等。存储器是能够用于携带或存储具有指令或数据结构形式的期望的程序代码并能够由计算机存取的任何其他介质，但不限于此。本申请实施例中的存储器702还可以是电路或者其它任意能够实现存储功能的装置，用于存储程序指令和/或数据。
本申请实施例提供了一种计算机可读存储介质，用于储存为上述电子设备所用的计算机可读指令，计算机可读指令被处理器执行时，使得处理器执行上述图像处理方法的步骤。
在一个实施例中，提供了一种计算机程序产品或计算机可读指令，该计算机程序产品或计算机可读指令包括计算机可读指令，该计算机可读指令存储在计算机可读存储介质中。计算机设备的处理器从计算机可读存储介质读取该计算机可读指令，处理器执行该计算机可读指令，使得该计算机设备执行上述各方法实施例中的步骤。
上述计算机可读存储介质可以是计算机能够存取的任何可用介质或数据存储设备，包括但不限于磁性存储器(例如软盘、硬盘、磁带、磁光盘(MO)等)、光学存储器(例如CD、DVD、BD、HVD等)、以及半导体存储器(例如ROM、EPROM、EEPROM、非易失性存储器(NAND FLASH)、固态硬盘(SSD))等。
以上，以上实施例仅用以对本申请的技术方案进行了详细介绍，但以上实施例的说明只是用于帮助理解本申请实施例的方法，不应理解为对本申请实施例的限制。本技术领域的技术人员可轻易想到的变化或替换，都应涵盖在本申请实施例的保护范围之内。

Claims (20)

  1. 一种图像处理方法，由电子设备执行，所述方法包括：
    根据原图像中颜色特征值落入各个颜色特征区间内的像素点的数量，获得所述原图像对应的第一直方图；
    根据所述原图像的上采样图像中颜色特征值落入所述各个颜色特征区间内的像素点的数量，获得所述上采样图像对应的第二直方图；
    根据所述第一直方图和所述第二直方图，分别确定出所述第二直方图中各个颜色特征区间所匹配的目标颜色特征区间；及
    根据所述第二直方图中各个颜色特征区间所匹配的目标颜色特征区间，对所述上采样图像中落入所述各个颜色特征区间的像素点的颜色特征值进行处理，获得目标图像。
  2. 根据权利要求1所述的方法，其中所述根据所述第一直方图和所述第二直方图，分别确定出所述第二直方图中各个颜色特征区间所匹配的目标颜色特征区间包括：
    针对所述第二直方图中的任一颜色特征区间，确定所述第一直方图中各个颜色特征区间对应的占比值和所述任一颜色特征区间对应的占比值之间的差值，从满足指定条件的差值在所述第一直方图中对应的颜色特征区间中，确定出所述任一颜色特征区间所匹配的目标颜色特征区间，其中，所述第一直方图中的每个颜色特征区间对应的占比值为：所述原图像中颜色特征值落入对应的颜色特征区间内的像素点的数量与所述原图像包含的像素点总数的比值，所述第二直方图中的每个颜色特征区间对应的占比值为：所述上采样图像中颜色特征值落入对应的颜色特征区间内的像素点的数量与所述上采样图像包含的像素点总数的比值。
  3. 根据权利要求2所述的方法，其中若存在至少两个差值满足所述指定条件，则所述从满足指定条件的差值在所述第一直方图中对应的颜色特征区间中，确定出所述任一颜色特征区间所匹配的目标颜色特征区间包括：
    确定所述至少两个差值中每个差值在所述第一直方图中对应的颜色特征区间；及
    从确定出的至少两个颜色特征区间中选出与所述任一颜色特征区间的等级差最小的颜色特征区间，确定为所述任一颜色特征区间所匹配的目标颜色特征区间。
  4. 根据权利要求2所述的方法，其中若仅存在一个满足所述指定条件的差值，则所述从满足指定条件的差值在所述第一直方图中对应的颜色特征区间中，确定出所述任一颜色特征区间所匹配的目标颜色特征区间包括：
    将所述满足所述指定条件的差值在第一直方图中对应的颜色特征区间确定为所述任一颜色特征区间所匹配的目标颜色特征区间。
  5. 根据权利要求1所述的方法，其中所述根据所述第一直方图和所述第二直方图，分别确定出所述第二直方图中各个颜色特征区间所匹配的目标颜色特征区间包括：
    根据所述第一直方图，确定所述原图像中像素点的颜色特征区间对应的第一均值和第一方差；
    根据所述第二直方图，确定所述上采样图像中像素点的颜色特征区间对应的第二均值和第二方差；
    根据所述第一均值、所述第一方差、所述第二均值和所述第二方差，确定所述第一直方图中的颜色特征区间和所述第二直方图中的颜色特征区间之间的第一映射关系；及
    根据所述第一映射关系，确定所述第二直方图中各个颜色特征区间所匹配的目标颜色特征区间。
  6. 根据权利要求1至5任一所述的方法，其中所述根据所述第二直方图中各个颜色特征区间所匹配的目标颜色特征区间，对所述上采样图像中落入所述各个颜色特征区间的像素点的颜色特征值进行处理，获得目标图像包括：
    若每个颜色特征区间对应一个颜色特征值，则分别将所述上采样图像中 落入所述各个颜色特征区间的像素点的颜色特征值，调整为所述各个颜色特征区间所匹配的目标颜色特征区间对应的颜色特征值，获得目标图像；及
    若每个颜色特征区间对应一个颜色特征值范围，则分别确定所述第二直方图中各个颜色特征区间对应的颜色特征值范围和各个颜色特征区间所匹配的目标颜色特征区间对应的颜色特征值范围之间的第二映射关系，将所述上采样图像中落入所述各个颜色特征区间的像素点的颜色特征值调整为根据所述第二映射关系确定的颜色特征值。
  7. 根据权利要求1至5任一所述的方法，其中所述颜色特征值包括以下至少一种：像素点的灰度值、像素点的亮度值、像素点的颜色值。
  8. 一种图像处理装置，包括：
    第一统计模块，用于根据原图像中颜色特征值落入各个颜色特征区间内的像素点的数量，获得所述原图像对应的第一直方图；
    第二统计模块，用于根据所述原图像的上采样图像中颜色特征值落入所述各个颜色特征区间内的像素点的数量，获得所述上采样图像对应的第二直方图；
    匹配模块，用于根据所述第一直方图和所述第二直方图，分别确定出所述第二直方图中各个颜色特征区间所匹配的目标颜色特征区间；及
    处理模块，用于根据所述第二直方图中各个颜色特征区间所匹配的目标颜色特征区间，对所述上采样图像中落入所述各个颜色特征区间的像素点的颜色特征值进行处理，获得目标图像。
  9. 根据权利要求8所述的装置，其中所述匹配模块，具体用于针对所述第二直方图中的任一颜色特征区间，确定所述第一直方图中各个颜色特征区间对应的占比值和所述任一颜色特征区间对应的占比值之间的差值，从满足指定条件的差值在所述第一直方图中对应的颜色特征区间中，确定出所述任一颜色特征区间所述匹配的目标颜色特征区间，其中，所述第一直方图中的每个颜色特征区间对应的占比值为：所述原图像中颜色特征值落入对应的颜 色特征区间内的像素点的数量与所述原图像包含的像素点总数的比值，所述第二直方图中的每个颜色特征区间对应的占比值为：所述上采样图像中颜色特征值落入对应的颜色特征区间内的像素点的数量与所述上采样图像包含的像素点总数的比值。
  10. 根据权利要求9所述的装置，其中所述匹配模块，具体用于若存在至少两个差值满足所述指定条件，则确定所述至少两个差值中每个差值在所述第一直方图中对应的颜色特征区间，从确定出的至少两个颜色特征区间中选出与所述任一颜色特征区间的等级差最小的颜色特征区间，确定为所述任一颜色特征区间所述匹配的目标颜色特征区间。
  11. 根据权利要求9所述的装置，其中所述匹配模块，具体用于若仅存在一个满足所述指定条件的差值，则将所述满足所述指定条件的差值在第一直方图中对应的颜色特征区间确定为所述任一颜色特征区间所匹配的目标颜色特征区间。
  12. 根据权利要求8所述的装置，其中所述匹配模块，具体用于：
    根据所述第一直方图，确定所述原图像中像素点的颜色特征区间对应的第一均值和第一方差；
    根据所述第二直方图，确定所述上采样图像中像素点的颜色特征区间对应的第二均值和第二方差；
    根据所述第一均值、所述第一方差、所述第二均值和所述第二方差，确定所述第一直方图中的颜色特征区间和所述第二直方图中的颜色特征区间之间的第一映射关系；及
    根据所述第一映射关系，确定所述第二直方图中各个颜色特征区间所匹配的目标颜色特征区间。
  13. 根据权利要求8至12任一所述的装置，其中所述处理模块，具体用于：
    若每个颜色特征区间对应一个颜色特征值，则分别将所述上采样图像中 落入所述各个颜色特征区间的像素点的颜色特征值，调整为所述各个颜色特征区间所匹配的目标颜色特征区间对应的颜色特征值，获得目标图像；及
    若每个颜色特征区间对应一个颜色特征值范围，则分别确定所述第二直方图中各个颜色特征区间对应的颜色特征值范围和各个颜色特征区间所匹配的目标颜色特征区间对应的颜色特征值范围之间的第二映射关系，将所述上采样图像中落入所述各个颜色特征区间的像素点的颜色特征值调整为根据所述第二映射关系确定的颜色特征值。
  14. 根据权利要求8至12任一所述的装置，其中所述颜色特征值包括以下至少一种：像素点的灰度值、像素点的亮度值、像素点的颜色值。
  15. 一种电子设备，包括存储器和一个或多个处理器，存储器中储存有计算机可读指令，计算机可读指令被处理器执行时，使得一个或多个处理器执行以下步骤：
    根据原图像中颜色特征值落入各个颜色特征区间内的像素点的数量，获得所述原图像对应的第一直方图；
    根据所述原图像的上采样图像中颜色特征值落入所述各个颜色特征区间内的像素点的数量，获得所述上采样图像对应的第二直方图；
    根据所述第一直方图和所述第二直方图，分别确定出所述第二直方图中各个颜色特征区间所匹配的目标颜色特征区间；及
    根据所述第二直方图中各个颜色特征区间所匹配的目标颜色特征区间，对所述上采样图像中落入所述各个颜色特征区间的像素点的颜色特征值进行处理，获得目标图像。
  16. 根据权利要求15所述的电子设备，其中所述计算机可读指令被处理器执行时，使得一个或多个处理器还执行以下步骤：
    针对所述第二直方图中的任一颜色特征区间，确定所述第一直方图中各个颜色特征区间对应的占比值和所述任一颜色特征区间对应的占比值之间的差值，从满足指定条件的差值在所述第一直方图中对应的颜色特征区间中， 确定出所述任一颜色特征区间所匹配的目标颜色特征区间，其中，所述第一直方图中的每个颜色特征区间对应的占比值为：所述原图像中颜色特征值落入对应的颜色特征区间内的像素点的数量与所述原图像包含的像素点总数的比值，所述第二直方图中的每个颜色特征区间对应的占比值为：所述上采样图像中颜色特征值落入对应的颜色特征区间内的像素点的数量与所述上采样图像包含的像素点总数的比值。
  17. 根据权利要求16所述的电子设备，其中所述计算机可读指令被处理器执行时，使得一个或多个处理器还执行以下步骤：
    若存在至少两个差值满足所述指定条件，则确定所述至少两个差值中每个差值在所述第一直方图中对应的颜色特征区间；及
    从确定出的至少两个颜色特征区间中选出与所述任一颜色特征区间的等级差最小的颜色特征区间，确定为所述任一颜色特征区间所匹配的目标颜色特征区间。
  18. 根据权利要求16所述的电子设备，其中所述计算机可读指令被处理器执行时，使得一个或多个处理器还执行以下步骤：
    若仅存在一个满足所述指定条件的差值，则将所述满足所述指定条件的差值在第一直方图中对应的颜色特征区间确定为所述任一颜色特征区间所匹配的目标颜色特征区间。
  19. 根据权利要求15所述的电子设备，其中所述计算机可读指令被处理器执行时，使得一个或多个处理器还执行以下步骤：
    根据所述第一直方图，确定所述原图像中像素点的颜色特征区间对应的第一均值和第一方差；
    根据所述第二直方图，确定所述上采样图像中像素点的颜色特征区间对应的第二均值和第二方差；
    根据所述第一均值、所述第一方差、所述第二均值和所述第二方差，确定所述第一直方图中的颜色特征区间和所述第二直方图中的颜色特征区间之 间的第一映射关系；及
    根据所述第一映射关系，确定所述第二直方图中各个颜色特征区间所匹配的目标颜色特征区间。
  20. 一种存储有计算机可读指令的非易失性存储介质，所述计算机可读指令被处理器执行时，使得所述处理器执行如权利要求1至7任一项所述方法。
PCT/CN2020/122203 2020-01-14 2020-10-20 图像处理方法、装置、电子设备及存储介质 Ceased WO2021143241A1 (zh)

Priority Applications (4)

Application Number Priority Date Filing Date Title
EP20913353.7A EP4024327A4 (en) 2020-01-14 2020-10-20 IMAGE PROCESSING METHOD, EQUIPMENT, ELECTRONIC DEVICE AND STORAGE MEDIA
KR1020227006382A KR102617626B1 (ko) 2020-01-14 2020-10-20 이미지 처리 방법, 장치, 전자 디바이스, 및 저장 매체
JP2022520267A JP7508135B2 (ja) 2020-01-14 2020-10-20 画像処理方法、画像処理装置、電子機器、及びコンピュータプログラム
US17/573,491 US12260521B2 (en) 2020-01-14 2022-01-11 Image processing method and apparatus, electronic device, and storage medium

Applications Claiming Priority (2)

Application Number Priority Date Filing Date Title
CN202010038997.5 2020-01-14
CN202010038997.5A CN111260593B (zh) 2020-01-14 2020-01-14 图像处理方法、装置、电子设备及存储介质

Related Child Applications (1)

Application Number Title Priority Date Filing Date
US17/573,491 Continuation US12260521B2 (en) 2020-01-14 2022-01-11 Image processing method and apparatus, electronic device, and storage medium

Publications (1)

Publication Number Publication Date
WO2021143241A1 true WO2021143241A1 (zh) 2021-07-22

Family

ID=70945241

Family Applications (1)

Application Number Title Priority Date Filing Date
PCT/CN2020/122203 Ceased WO2021143241A1 (zh) 2020-01-14 2020-10-20 图像处理方法、装置、电子设备及存储介质

Country Status (6)

Country Link
US (1) US12260521B2 (zh)
EP (1) EP4024327A4 (zh)
JP (1) JP7508135B2 (zh)
KR (1) KR102617626B1 (zh)
CN (1) CN111260593B (zh)
WO (1) WO2021143241A1 (zh)

Cited By (1)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN116310448A (zh) * 2023-05-24 2023-06-23 山东曙岳车辆有限公司 基于计算机视觉的集装箱拼装匹配性检测方法

Families Citing this family (14)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN111260593B (zh) * 2020-01-14 2023-03-14 腾讯科技（深圳）有限公司 图像处理方法、装置、电子设备及存储介质
CN113160038B (zh) * 2021-04-28 2024-03-26 北京达佳互联信息技术有限公司 一种图像风格迁移方法、装置、电子设备及存储介质
CN113129392B (zh) * 2021-05-17 2022-07-26 杭州万事利丝绸文化股份有限公司 一种颜色搭配方法及系统
CN113421209B (zh) * 2021-06-21 2022-12-30 安谋科技（中国）有限公司 图像处理方法、片上系统、电子设备和介质
CN113538280B (zh) * 2021-07-20 2022-03-04 江苏天汇空间信息研究院有限公司 基于矩阵二值化的遥感影像去除彩线和残缺影像拼接方法
US20250104186A1 (en) * 2021-10-29 2025-03-27 Nvidia Corporation Image upsampling using one or more neural networks
CN113933294B (zh) * 2021-11-08 2023-07-18 中国联合网络通信集团有限公司 浓度检测方法及装置
US12493990B2 (en) * 2022-03-17 2025-12-09 Advanced Micro Devices, Inc. Locking mechanism for image classification
CN115082703B (zh) * 2022-07-19 2022-11-11 深圳大学 概念关联颜色提取方法、装置、计算机设备和存储介质
CN115713919B (zh) * 2022-11-24 2026-01-02 北京奕斯伟计算技术股份有限公司 背光调整方法、装置、终端及存储介质
CN117115838B (zh) * 2023-07-07 2026-03-27 南京费莎医疗科技有限公司 朱墨时序分析检测方法
CN117765276B (zh) * 2023-12-14 2025-07-11 成都信息工程大学 无人机空中视角人员识别方法、介质和电子设备
CN119205939A (zh) * 2024-09-04 2024-12-27 江苏省家禽科学研究所 一种基于机器视觉的鸡肤色评定方法及系统
CN120451803B (zh) * 2025-07-10 2025-09-12 四川省地质调查研究院测绘地理信息中心 一种基于图像识别的自然资源调查成果分析方法及系统

Citations (7)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN103778900A (zh) * 2012-10-23 2014-05-07 浙江大华技术股份有限公司 一种图像处理方法及系统
US20180061028A1 (en) * 2016-08-31 2018-03-01 Adobe Systems Incorporated Image lighting transfer via multi-dimensional histogram matching
CN108090887A (zh) * 2016-11-23 2018-05-29 杭州海康威视数字技术股份有限公司 一种视频图像处理方法及装置
CN109003249A (zh) * 2017-06-07 2018-12-14 展讯通信(天津)有限公司 增强图像细节的方法、装置及终端
CN109064434A (zh) * 2018-06-28 2018-12-21 广州视源电子科技股份有限公司 图像增强的方法、装置、存储介质及计算机设备
CN110049332A (zh) * 2019-04-11 2019-07-23 深圳市朗驰欣创科技股份有限公司 一种图像压缩方法、图像压缩装置及电子设备
CN111260593A (zh) * 2020-01-14 2020-06-09 腾讯科技（深圳）有限公司 图像处理方法、装置、电子设备及存储介质

Family Cites Families (11)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
JP4672680B2 (ja) * 2007-02-05 2011-04-20 日本電信電話株式会社 画像処理方法、画像処理装置、画像処理プログラム及びそのプログラムを記録したコンピュータ読み取り可能な記録媒体
FR2947082B1 (fr) * 2009-06-22 2014-12-12 St Ericsson France Sas Procede et dispositif de traitement d'une image numerique pour eclaircir ladite image.
CN103460243B (zh) * 2011-03-17 2016-08-17 杜比实验室特许公司 使用直方图产生图像内容的替代版本
US9076236B2 (en) * 2013-09-12 2015-07-07 At&T Intellectual Property I, L.P. Guided image upsampling using bitmap tracing
WO2016188549A1 (en) 2015-05-22 2016-12-01 Huawei Technologies Co., Ltd. Apparatus and method for color calibration
EP3507773A1 (en) * 2016-09-02 2019-07-10 Artomatix Ltd. Systems and methods for providing convolutional neural network based image synthesis using stable and controllable parametric models, a multiscale synthesis framework and novel network architectures
JP2019016117A (ja) * 2017-07-05 2019-01-31 日本電信電話株式会社 画像調整装置、局所的コントラスト量計算装置、方法、及びプログラム
CN107492070B (zh) * 2017-07-10 2019-12-03 华北电力大学 一种双通道卷积神经网络的单图像超分辨率计算方法
CN108924525B (zh) * 2018-06-06 2021-07-06 平安科技（深圳）有限公司 图像亮度调整方法、装置、计算机设备及存储介质
CN108900819B (zh) * 2018-08-20 2020-09-15 Oppo广东移动通信有限公司 图像处理方法、装置、存储介质及电子设备
JP2021018123A (ja) * 2019-07-19 2021-02-15 ソニーセミコンダクタソリューションズ株式会社 信号処理装置、信号処理方法

Patent Citations (7)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN103778900A (zh) * 2012-10-23 2014-05-07 浙江大华技术股份有限公司 一种图像处理方法及系统
US20180061028A1 (en) * 2016-08-31 2018-03-01 Adobe Systems Incorporated Image lighting transfer via multi-dimensional histogram matching
CN108090887A (zh) * 2016-11-23 2018-05-29 杭州海康威视数字技术股份有限公司 一种视频图像处理方法及装置
CN109003249A (zh) * 2017-06-07 2018-12-14 展讯通信(天津)有限公司 增强图像细节的方法、装置及终端
CN109064434A (zh) * 2018-06-28 2018-12-21 广州视源电子科技股份有限公司 图像增强的方法、装置、存储介质及计算机设备
CN110049332A (zh) * 2019-04-11 2019-07-23 深圳市朗驰欣创科技股份有限公司 一种图像压缩方法、图像压缩装置及电子设备
CN111260593A (zh) * 2020-01-14 2020-06-09 腾讯科技（深圳）有限公司 图像处理方法、装置、电子设备及存储介质

Non-Patent Citations (1)

* Cited by examiner, † Cited by third party
Title
See also references of EP4024327A4 *

Cited By (1)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN116310448A (zh) * 2023-05-24 2023-06-23 山东曙岳车辆有限公司 基于计算机视觉的集装箱拼装匹配性检测方法

Also Published As

Publication number Publication date
JP2022550565A (ja) 2022-12-02
KR102617626B1 (ko) 2023-12-27
JP7508135B2 (ja) 2024-07-01
US20220138920A1 (en) 2022-05-05
US12260521B2 (en) 2025-03-25
EP4024327A1 (en) 2022-07-06
CN111260593A (zh) 2020-06-09
KR20220043158A (ko) 2022-04-05
EP4024327A4 (en) 2023-01-04
CN111260593B (zh) 2023-03-14

Similar Documents

Publication Publication Date Title
CN111260593B (zh) 图像处理方法、装置、电子设备及存储介质
US11030729B2 (en) Image processing method and apparatus for adjusting a dynamic range of an image
CN111163268B (zh) 一种图像处理方法、装置以及计算机存储介质
WO2021093718A1 (zh) 视频处理方法、视频修复方法、装置及设备
CN110136071A (zh) 一种图像处理方法、装置、电子设备及存储介质
WO2021223205A1 (en) Encoder, decoder, system, and method for determining tone mapping curve parameters
EP4372730B1 (en) Image enhancement method and apparatus, computer device, and storage medium
CN103826113B (zh) 一种色彩还原方法及装置
CN105574896B (zh) 一种面向高分辨率视频的高效背景建模方法
CN108694030B (zh) 处理高动态范围图像的方法和装置
CN109934783B (zh) 图像处理方法、装置、计算机设备和存储介质
CN115471413A (zh) 图像处理方法及装置、计算机可读存储介质和电子设备
JP7780629B2 (ja) トーンマッピング方法およびシステム、ならびにデバイス
CN114998122A (zh) 一种低照度图像增强方法
US8331695B1 (en) Integrated circuit having a circuit for and method of updating parameters associated with a background estimation portion of a video frame
CN115442575A (zh) 一种基于视频质量评估的视频自适应传输方法及系统
US9832395B2 (en) Information processing method applied to an electronic device and electronic device having at least two image capturing units that have the same image capturing direction
WO2016197323A1 (zh) 视频编解码方法和视频编解码器
CN113613024B (zh) 视频预处理方法及设备
HK40024840B (zh) 图像处理方法、装置、电子设备及存储介质
HK40024840A (zh) 图像处理方法、装置、电子设备及存储介质
CN117793510A (zh) 摄像机补光灯控制方法、装置、计算机设备以及存储介质
CN115942002A (zh) 视频传输方法、装置、终端及存储介质
CN114385847B (zh) 一种图片数据处理方法、装置、计算机设备及存储介质
CN116320776A (zh) 图像处理方法、装置、芯片、电子设备和存储介质

Legal Events

Date Code Title Description
121 Ep: the epo has been informed by wipo that ep was designated in this application

Ref document number: 20913353

Country of ref document: EP

Kind code of ref document: A1

ENP Entry into the national phase

Ref document number: 20227006382

Country of ref document: KR

Kind code of ref document: A

ENP Entry into the national phase

Ref document number: 2022520267

Country of ref document: JP

Kind code of ref document: A

ENP Entry into the national phase

Ref document number: 2020913353

Country of ref document: EP

Effective date: 20220330

NENP Non-entry into the national phase

Ref country code: DE