WO2019157966A1 - 图像增强方法、数据处理设备及存储介质 - Google Patents

图像增强方法、数据处理设备及存储介质 Download PDF

Info

Publication number
WO2019157966A1
WO2019157966A1 PCT/CN2019/074067 CN2019074067W WO2019157966A1 WO 2019157966 A1 WO2019157966 A1 WO 2019157966A1 CN 2019074067 W CN2019074067 W CN 2019074067W WO 2019157966 A1 WO2019157966 A1 WO 2019157966A1
Authority
WO
WIPO (PCT)
Prior art keywords
window
image
processed image
pixel
determining
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/CN2019/074067
Other languages
English (en)
French (fr)
Inventor
赵永飞
郭治姣
汪亮
翟海昌
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 EP19754702.9A priority Critical patent/EP3754590B1/en
Publication of WO2019157966A1 publication Critical patent/WO2019157966A1/zh
Priority to US16/902,481 priority patent/US11263727B2/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
    • G06T5/70—Denoising; Smoothing
    • G—PHYSICS
    • G06—COMPUTING OR CALCULATING; COUNTING
    • G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
    • G06T5/00—Image enhancement or restoration
    • G06T5/73—Deblurring; Sharpening
    • G—PHYSICS
    • G06—COMPUTING OR CALCULATING; COUNTING
    • G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
    • G06T5/00—Image enhancement or restoration
    • G06T5/50—Image enhancement or restoration using two or more images, e.g. averaging or subtraction
    • G—PHYSICS
    • G06—COMPUTING OR CALCULATING; COUNTING
    • G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
    • G06T5/00—Image enhancement or restoration
    • G06T5/90—Dynamic range modification of images or parts thereof
    • G06T5/92—Dynamic range modification of images or parts thereof based on global image properties
    • 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
    • G06T2207/10016—Video; Image sequence
    • 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
    • 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
    • G06T2207/20—Special algorithmic details
    • G06T2207/20076—Probabilistic image processing
    • 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
    • G06T2207/20—Special algorithmic details
    • G06T2207/20172—Image enhancement details
    • G06T2207/20192—Edge enhancement; Edge preservation
    • 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
    • G06T2207/20—Special algorithmic details
    • G06T2207/20212—Image combination
    • G06T2207/20224—Image subtraction

Definitions

  • the present application relates to the field of image processing technologies, and in particular, to an image enhancement method, a data processing device, and a storage medium.
  • the present application proposes an image enhancement scheme to improve the image enhancement effect.
  • an image enhancement method comprising: performing a edge-preserving filtering process on an original image to obtain a first processed image; acquiring a detail feature of the original image; and according to the detail feature and the first process The image determines a second processed image; the original image is used as a guiding image, and the second processed image is processed based on a directed image filtering manner to obtain a third processed image.
  • an image enhancement apparatus including: a first smoothing unit configured to perform edge-preserving filtering processing on an original image to obtain a first processed image; and a detail acquiring unit configured to acquire the original image a detail feature; an image superimposing unit, configured to determine a second processed image according to the detail feature and the first processed image; an image enhancement unit, configured to use the original image as a guide image, based on a guide image filtering manner The second processed image is processed to obtain a third processed image.
  • a video service system comprising: an image enhancement device according to the present application.
  • a terminal device comprising an image enhancement device according to the present application.
  • a data processing apparatus comprising: one or more processors, a memory, and one or more programs.
  • One or more programs are stored in the memory and configured to be executed by the one or more processors, the one or more programs including instructions for performing the image enhancement method of the present application.
  • a storage medium storing one or more programs, the one or more programs comprising instructions that, when executed by a data processing device, cause the data processing device to execute Image enhancement method applied for.
  • the image processing scheme can acquire a third processed image having a high degree of smoothness and enhanced detail features.
  • the image enhancement scheme can improve the image noise reduction degree and enhance the detail feature, thereby making the picture more clear, the overall brightness more uniform, the layering feeling stronger, the contrast is higher, and the color is more vivid.
  • FIG. 1A shows a schematic diagram of an application scenario according to some embodiments of the present application
  • FIG. 1B shows a schematic diagram of an application scenario according to some embodiments of the present application
  • FIG. 2A shows a flowchart of an image enhancement method 200 in accordance with some embodiments of the present application
  • FIG. 2B and 2C illustrate an original image and a first processed image, respectively, in accordance with some embodiments of the present application
  • FIG. 2D illustrates a second processed image in accordance with some embodiments of the present application
  • Figure 2F shows a comparison of zone Z2 in Figure 2E with Z1 in Figure 2B;
  • FIG. 3 illustrates a flow diagram of a method 300 of acquiring a third processed image, in accordance with some embodiments of the present application
  • FIG. 4 illustrates a flow diagram of a method 400 of determining a slope of a first linear relationship, in accordance with some embodiments of the present application
  • FIG. 5 illustrates a flow diagram of a method 500 of determining a slope of a first linear relationship, in accordance with some embodiments of the present application
  • FIG. 6 shows a flowchart of a method 600 of determining a weight value, in accordance with some embodiments of the present application
  • FIG. 7 illustrates a flow diagram of an image enhancement method 700 in accordance with some embodiments of the present application.
  • FIG. 8 shows a schematic diagram of an image enhancement device 800 in accordance with some embodiments of the present application.
  • FIG. 9 shows a schematic diagram of an image enhancement device 900 in accordance with some embodiments of the present application.
  • Figure 10 is a diagram showing the composition of a data processing device.
  • FIG. 1A shows a schematic diagram of an application scenario in accordance with some embodiments of the present application.
  • the application scenario includes a terminal device 110 and a video service system 120.
  • the terminal device 110 may be, for example, various devices such as a desktop computer, a mobile phone, a tablet computer, a smart TV, or a laptop computer.
  • the terminal device 110 can include a video application 111.
  • the terminal device 110 can download and install the video application 111.
  • the video application 111 can also be pre-installed on the terminal device 110.
  • Video service system 120 may, for example, include one or more servers. In some embodiments, video services system 120 can store video content.
  • Video application 111 may obtain video content from video services system 120 over network 130.
  • Video application 111 may perform video enhancement processing on video content from local or from video services system 120.
  • the video application 111 can be, for example, a video playback client (such as a Tencent video client or an iQiyi video client), a social application, or a browser, and the like, and is not limited in this application.
  • FIG. 1B shows a schematic diagram of an application scenario in accordance with some embodiments of the present application.
  • the application scenario includes a terminal device 140 and a video service system 150.
  • Video service system 150 can include a video processing application 151.
  • the video processing application 151 can perform image enhancement processing on the video.
  • the terminal device 140 can include a video playback application 141.
  • the video playing application 141 is, for example, a variety of applications including a video playing function such as a video playing client, a social application, or a browser.
  • the video playback application 141 can retrieve and play the image-enhanced video from the video service system 150.
  • both the terminal device 110 and the video service system 150 can be used to enhance the video.
  • various devices of the video enhancement process can perform image enhancement processing on each image frame in the video.
  • the image enhancement mode according to the present application will be described in detail below with reference to FIG.
  • FIG. 2A shows a flow chart of an image enhancement method 200 in accordance with some embodiments of the present application.
  • the image enhancement method 200 is performed, for example, by a data processing device such as the terminal device 110 or the video service system 150.
  • the image enhancement method 200 may include a step S201 of performing an edge-preserving filtering process on the original image to obtain a first processed image.
  • step S201 uses the original image as a navigation map, and processes the original image based on a Guided Image Filtering method to obtain a first processed image.
  • step S201 may be processed by various edge-preserving filtering methods such as The median filtering, Bilateral Filtering, or Weighted Least Square Filtering.
  • step S201 can perform denoising processing on the original image.
  • Figures 2B and 2C show the original image and the first processed image, respectively, in some embodiments.
  • the first processed image has a higher smoothness than the original image.
  • step S202 the detail features of the original image are acquired.
  • the detail features may include, for example, edge features and texture features, and the like.
  • step S202 may first determine a color difference between the original image and the first processed image, and then determine the detail feature based on the color difference.
  • step S202 may determine a difference between the original image and the first processed image on one or more color channels and use the difference on the one or more color channels as a color difference.
  • step S202 may determine a grayscale value difference between the original image and the first processed image as a color difference.
  • the original image can include multiple color channels.
  • Step S202 may determine a difference between the original image and the first processed image on the plurality of color channels, and use the difference as a color difference.
  • the original image is in red, green and blue (RGB) format.
  • the color difference acquired in step S202 may include a difference of a red channel, a difference of a green channel, and a difference of a blue channel.
  • the color difference determined in step S202 may include, for example, feature information such as edges and textures in the original image.
  • step S202 may determine the detail feature of the original image based on the color difference. For example, step S202 can directly take the color difference as a detail feature. For another example, step S202 may enhance the color difference according to the adjustment coefficient to obtain an enhanced color difference, and use the enhanced color difference as the detail feature corresponding to the original image. Here, the larger the adjustment coefficient value, the higher the degree of enhancement. In addition, step S202 may also acquire the detailed features of the original image in other suitable manners.
  • step S203 a second processed image is determined based on the detail feature and the first processed image.
  • step S203 may superimpose the detail feature on the first processed image to obtain the second processed image.
  • step S203 may also adopt other manners to determine the second processed image using the detail feature and the first processed image.
  • step S203 may cause the second processed image to maintain a smoothness (ie, a degree of noise reduction) close to the first processed image, or may cause the second processed image to be made.
  • the details in the second processed image are more easily distinguished.
  • FIG. 2D shows the second processed image obtained by the processing of step S202 in the first processed image in FIG. 2C. Compared to Figure 2C, the second processed image detail in Figure 2D is more easily distinguished.
  • step S204 the original image is used as a guide image, and the second processed image is processed based on the map filtering method to obtain a third processed image.
  • step S204 can perform noise reduction on the second processed image to increase the degree of noise reduction of the third processed image.
  • the detail feature in the second processed image can be retained in accordance with the detailed feature pattern of the original image.
  • step S204 performs image enhancement processing on the second processed image by using the original image as the guide image, so that the detailed features in the third processed image are more consistent with the overall overview of the original image.
  • step S204 can make the detail features in the third processed image more toward the overall overview of the original image, that is, the details are more integrated with the overall image.
  • FIG. 2E shows a third processed image corresponding to the second processed image in FIG. 2D.
  • Figure 2F shows a comparison of zone Z2 in Figure 2E with Z1 in Figure 2B. The contents of the area Z1 and the area Z2 are the same. Compared with the Z1 representing the original image, the Z2 picture representing the third processed image is clearer, the overall brightness is more uniform, the layering is stronger, and the contrast is higher.
  • Figures 2D-2F are shown as grayscale images, and may actually be color images. Step S204 can also make the color of the third processed image of the color more vivid.
  • the image enhancement method 200 can acquire a third processed image having a high degree of smoothness and enhanced detail features by a combination of the above steps.
  • the image enhancement method 200 can improve the image noise reduction degree and enhance the detail feature, thereby making the picture more clear, the overall brightness more uniform, the layering feeling stronger, the contrast higher, and the color more vivid.
  • the higher the contrast the higher the color gradation of the image, and the more full the picture.
  • the image enhancement method 200 can achieve adaptive processing of image content through the combination of the above steps, thereby improving the robustness of image processing.
  • the image enhancement method 200 can make the brightness of the third processed image more uniform by adopting the guide pattern filtering method.
  • FIG. 3 illustrates a flow diagram of a method 300 of acquiring a third processed image, in accordance with some embodiments of the present application.
  • the pixel point in the guide image is referred to as a first pixel point.
  • a pixel in the second processed image is referred to as a second pixel, and a pixel in the third processed image is referred to as a third pixel.
  • the method 300 for acquiring a third processed image may include step S301, for each first pixel in the navigation map, determining a corresponding position in each of the first window and the third processed image including the first pixel
  • the first linear relationship between the second windows is used to describe a linear relationship between each first pixel in the first window and a corresponding pixel in the corresponding second window.
  • the size of each first window and the size of the second window of the corresponding position are both the first size.
  • the first size is used to describe the side length of the window.
  • the first size can also be represented by a radius.
  • the radius is r and the first size can be expressed as (2r+1). Accordingly, the window satisfying the first size may contain (2r+1)*(2r+1) pixels.
  • the first linear relationship determined in step S301 can be expressed as the following formula:
  • w k represents the first window centered on the pixel point k in the orientation map
  • I i represents the ith first pixel point in the first window
  • q i represents a first position corresponding to the I i position in the third processed image
  • a k represents the slope of the first linear relationship
  • b k represents the intercept.
  • step S301 can determine the slope of the first linear relationship by method 400.
  • 4 shows a flow diagram of a method 400 of determining a slope of a first linear relationship, in accordance with some embodiments of the present application.
  • step S401 a variance corresponding to the first window is determined.
  • step S402 a covariance of the third window of the corresponding position in the first window and the second processed image is determined.
  • step S403 the sum of the variance and the regularization coefficient threshold is determined.
  • the regularization coefficient threshold is a very small positive number.
  • step S404 the ratio of the covariance obtained in step S402 to the sum obtained in step S403 is taken as the slope.
  • step S301 can determine the slope of the first linear relationship according to the following formula:
  • a k represents the slope
  • w k represents the first window centered on the pixel point k
  • I i represents the color value of the ith pixel point in the first window
  • p i represents the ith pixel point in the third window a color value
  • the third window is a window corresponding to the first window position in the second processed image
  • u k represents a color average of all the pixels in the first window
  • ⁇ represents the regularization coefficient threshold
  • represents the total number of pixels in the first window.
  • step S301 can determine the slope of the first linear relationship by method 500.
  • FIG. 5 illustrates a flow diagram of a method 500 of determining a slope of a first linear relationship, in accordance with some embodiments of the present application.
  • step S501 the variance corresponding to the first window is determined.
  • the variance corresponding to the first window refers to the variance of the color information of all the pixels in the first window.
  • step S502 a covariance of the third window of the corresponding position in the first window and the second processed image is determined.
  • step S503 a weight value of the detail feature of the first window in the first processed image is determined.
  • step S503 can be implemented as method 600.
  • FIG. 6 illustrates a flow diagram of a method 600 of determining a weight value, in accordance with some embodiments of the present application.
  • step S601 a first variance corresponding to each first pixel point in the navigation map and a corresponding second variance are determined.
  • the first variance corresponding to the first pixel point represents a variance corresponding to the window centered on the first pixel and having a size satisfying the second size.
  • a second variance corresponding to a first pixel represents a variance corresponding to a window centered at the first pixel and having a size that satisfies the first size.
  • the second size may, for example, represent the minimum window size used by the image processing process, for example, 3, that is, the radius is 1, but is not limited thereto.
  • step S602 a product of the first variance corresponding to the center point of the first window and the corresponding second variance is determined, and the binary one of the products is taken as the third value.
  • step S603 a product of the first variance corresponding to each first pixel point in the navigation map and the corresponding second variance is determined, and the first half of the product corresponding to each first pixel point is determined. The average value is taken as the fourth value.
  • the ratio of the third value to the fourth value is taken as the weight value.
  • step S504 a first ratio of the regularization coefficient threshold to the weight value is determined.
  • step S505 the sum of the covariance in step S502 and the first ratio in step S504 is taken as the first value, and the sum of the variance in step S502 and the first ratio in step S504 is taken as the second value, and the first value is The ratio of the second value is taken as the slope to be determined.
  • the method 500 determines the weight value and determines the slope using the weight value by using the second variance (ie, the variance of the color value in the window that satisfies the second size) in the process of determining the weight value.
  • the image enhancement method 200 when acquiring the third processed image using the second linear relationship associated with the slope, can cause the third processed image to always retain the strongest detail features in the original image.
  • the larger the detail gradient value the higher the detail intensity.
  • the detail features can be distinguished by the intensity of the detail.
  • the image enhancement method 200 can always retain the strongest detail features in the original image.
  • step S301 can determine the slope of the first linear relationship according to the following formula:
  • a k represents the slope
  • w k represents the first window centered on the pixel point k
  • I i represents the color value of the ith pixel point in the first window
  • p i represents the ith pixel point in the third window a color value
  • the third window is a window corresponding to the first window position in the second processed image
  • u k represents a color average of all pixels in the first window
  • ⁇ represents the regularization coefficient threshold
  • represents the total number of pixels in the first window
  • weight k represents the weight value of the first window.
  • step S301 can determine the weight value of the first window according to the following formula:
  • n represents the window radius of the first window
  • the variance of the color values of all the pixels in the window ⁇ i
  • r 1 2 represents the variance of the color values of all the pixels in the window centered on the i-th pixel in the first window and having a radius of 1
  • I represents a guide map
  • represents the total number of pixel points in the guide map.
  • Step S301 after determining a slope of a first linear relationship corresponding to each first window including a first pixel, may further determine an intercept of the first linear relationship according to the slope.
  • step S301 can determine the intercept according to the following formula:
  • a second linear relationship corresponding to the first pixel point is determined according to a first linear relationship corresponding to each first window.
  • the second linear relationship is used to describe a linear relationship between the first pixel and the corresponding pixel in the third processed image.
  • the slope of the second linear relationship is the mean of the slopes of the first linear relationship corresponding to the respective first windows.
  • step S303 the color value of each pixel in the third processed image is determined according to the color value of each first pixel in the navigation map and the second linear relationship corresponding to each first pixel, that is, the third processed image is acquired. .
  • step S303 can determine the color value of each pixel according to the following formula:
  • a k represents the slope
  • w k represents the first window centered on the pixel point k in the guidance map
  • I i represents the color value of the ith pixel point in the first window
  • represents the pixel point in the first window
  • b k represents the intercept of the first linear relationship corresponding to the first window
  • q i represents the color value of the i-th pixel in the second window corresponding to the first window position in the third processed image
  • FIG. 7 shows a flow diagram of an image enhancement method 700 in accordance with some embodiments of the present application.
  • the image enhancement method 700 can be performed, for example, by a data processing device such as the terminal device 110 or the video service system 150.
  • the image enhancement method 700 can include steps S701 through S704.
  • the implementation manners of steps S701 to S704 are consistent with steps S201-S204, and are not described herein again.
  • the image enhancement method 700 may further include a step S705 of performing a edge-preserving filtering process on the third processed image to obtain a fourth processed image.
  • step S705 uses the third processed image as a guide map, and processes the third processed image based on the map filtering method to obtain a fourth processed image. In this way, in step S705, by performing the edge-preserving filtering process on the third processed image itself as the guide image, the degree of noise reduction can be further improved, thereby improving the effect of image enhancement.
  • FIG. 8 shows a schematic diagram of an image enhancement device 800 in accordance with some embodiments of the present application.
  • the data processing device such as the terminal device 110 or the video service system 150 may include an image enhancement device 800.
  • the image enhancement device 800 may include a first smoothing unit 801, a detail acquisition unit 802, an image superimposition unit 803, and an image enhancement unit 804.
  • the first smoothing unit 801 is configured to perform a edge-preserving filtering process on the original image to obtain a first processed image.
  • the first smoothing unit 801 uses the original image as a guide map, and processes the original image based on the map filtering method to obtain a first processed image.
  • the detail obtaining unit 802 is configured to acquire a detail feature of the original image.
  • the detail acquisition unit 802 can determine a color difference between the original image and the first processed image, and determine the detail feature based on the color difference. In some embodiments, the detail acquisition unit 802 can determine the difference between the original image and the first processed image on one or more color channels and use the difference as a color difference. In some embodiments, the original image is a grayscale image. The detail acquisition unit 802 can determine a difference in gray value between the original image and the first processed image as a color difference. In other embodiments, the detail acquisition unit 802 can enhance the color difference in accordance with the adjustment factor to obtain an enhanced color difference and use the enhanced color difference as a detail feature. In other embodiments, the detail acquisition unit 802 can take a color difference as a detail feature.
  • the image superimposing unit 803 is configured to determine the second processed image according to the detail feature and the first processed image.
  • the image enhancement unit 804 uses the original image as a guide map, and processes the second processed image based on the map filtering method to obtain a third processed image.
  • the image enhancement unit 804 can determine between the first window including the first pixel and the second window of the corresponding position in the third processed image.
  • the first linear relationship wherein, the size of each first window and the size of the second window of the corresponding position are both the first size.
  • image enhancement unit 804 can determine the variance corresponding to the first window. Based on this, image enhancement unit 804 can determine the sum of the variance and the regularization coefficient threshold. Additionally, image enhancement unit 804 can also determine a covariance of the third window of the corresponding location in the first window and the second processed image. In this way, the image enhancement unit 804 can use the ratio of the covariance to the sum of the above as the slope of the first linear relationship corresponding to the first window. In some embodiments, image enhancement unit 804 can determine the slope of the first linear relationship according to the following formula:
  • a k represents the slope
  • w k represents the first window centered on the pixel point k
  • I i represents the color value of the ith pixel point in the first window
  • p i represents the third a color value of an i-th pixel in the window
  • the third window is a window corresponding to the first window position in the second processed image
  • u k represents a color mean of all pixels in the first window
  • ⁇ represents the regularization coefficient threshold
  • represents the total number of pixels in the first window.
  • the image enhancement unit 804 may also determine the intercept of the first linear relationship based on the slope.
  • image enhancement unit 804 can determine the variance corresponding to the first window.
  • the image enhancement unit 804 can determine a covariance of the second window of the corresponding location in the first window and the second processed image. Additionally, image enhancement unit 804 can determine a weight value for the detail feature of the first window in the first processed image.
  • the image enhancement unit 804 can determine a first variance corresponding to each of the first pixel points in the map and a corresponding second variance.
  • the first variance represents a variance of a window centered on the first pixel and having a size satisfying the second size.
  • the second variance represents the variance of the window centered at the first pixel and having a size that satisfies the first size. Based on this, the image enhancement unit 804 may determine a product of the first variance corresponding to the center point of the first window and the corresponding second variance, and use the binary one-time square as the third value.
  • the image enhancement unit 804 may determine a product of the first variance corresponding to each first pixel in the navigation map and the corresponding second variance, and divide the product corresponding to each first pixel by one-half The average value is taken as the fourth value.
  • the image enhancement unit 804 takes the ratio of the third value to the fourth value as the weight value.
  • image enhancement unit 804 can determine a first ratio of the regularization coefficient threshold to the weight value.
  • the image enhancement unit 804 may use the sum of the covariance and the first ratio as the first value, and the sum of the variance and the first ratio as the second value. In this way, the image enhancement unit 804 can use the ratio of the first value to the second value as the slope of the first linear relationship corresponding to the first window.
  • image enhancement unit 804 can determine the slope of the first linear relationship according to the following formula:
  • a k represents the slope
  • w k represents the first window centered on the pixel point k
  • I i represents the color value of the ith pixel point in the first window
  • p i represents the third window a color value of the i-th pixel
  • the third window is a window corresponding to the first window position in the second processed image
  • u k represents a color average of all pixels in the first window
  • Representing the color average of all pixels in the third window Representing the variance of the color values of all the pixels in the first window
  • ⁇ represents the regularization coefficient threshold
  • represents the total number of pixels in the first window
  • the weight k represents the weight value of the first window.
  • the image enhancement unit 804 can determine the weight value of the first window according to the following formula:
  • n a window radius of the first window
  • I represents the orientation map
  • the image enhancement unit 804 may determine a second linear relationship corresponding to the first pixel point according to the first linear relationship corresponding to each first window.
  • the second linear relationship is used to describe a linear relationship between the first pixel and the corresponding pixel in the third processed image.
  • the image enhancement unit 804 may determine the color value of each pixel in the third processed image according to the color value of each first pixel in the navigation map and the second linear relationship corresponding to each first pixel.
  • the image enhancement device 800 can acquire a third processed image having a high degree of smoothness and enhanced detail features.
  • the image enhancement device 800 can improve the image noise reduction degree and enhanced detail features, thereby making the picture more clear, the overall brightness more uniform, the layering feeling stronger, the contrast higher, and the color more vivid. Among them, the higher the contrast, the higher the color gradation of the image, and the more full the picture.
  • the image enhancement device 800 can implement adaptive processing of image content, thereby improving the robustness of image processing.
  • the image enhancement device 800 can make the brightness of the third processed image more uniform by adopting the guide pattern filtering method.
  • FIG. 9 shows a schematic diagram of an image enhancement device 900 in accordance with some embodiments of the present application.
  • the data processing device such as the terminal device 110 or the video service system 150 may include an image enhancement device 900.
  • the image enhancement device 900 may include a first smoothing unit 901, a detail acquisition unit 902, an image superimposition unit 903, and an image enhancement unit 904.
  • the embodiments of the first smoothing unit 901, the detail obtaining unit 902, the image superimposing unit 903, and the image enhancing unit 904 are identical to the first smoothing unit 801, the detail obtaining unit 802, the image superimposing unit 803, and the image enhancing unit 804, respectively. No longer.
  • the image enhancement device 900 may further include a second smoothing unit 905, configured to perform a edge-preserving filtering process on the third processed image to obtain a fourth processed image.
  • the second smoothing unit 905 processes the third processed image based on the map filtering method to obtain the fourth processed image.
  • the image enhancing device 900 can further increase the degree of noise reduction, thereby improving the effect of image enhancement.
  • FIG. 10 is a diagram showing the composition of a data processing device.
  • the data processing device can be implemented as a terminal device 110 or a video service system 150.
  • the data processing apparatus includes one or more processors (CPUs) 1002, communication modules 1004, memories 1006, user interfaces 1010, and a communication bus 1008 for interconnecting these components.
  • the processor 1002 can receive and transmit data through the communication module 1004 to effect network communication and/or local communication.
  • User interface 1010 includes one or more output devices 1012 that include one or more speakers and/or one or more visual displays. User interface 1010 also includes one or more input devices 1014. The user interface 1010 can receive, for example, an instruction of the remote controller, but is not limited thereto.
  • the memory 1006 may be a high speed random access memory such as DRAM, SRAM, DDR RAM, or other random access solid state storage device; or a nonvolatile memory such as one or more magnetic disk storage devices, optical disk storage devices, flash memory devices, Or other non-volatile solid-state storage devices.
  • a high speed random access memory such as DRAM, SRAM, DDR RAM, or other random access solid state storage device
  • nonvolatile memory such as one or more magnetic disk storage devices, optical disk storage devices, flash memory devices, Or other non-volatile solid-state storage devices.
  • the memory 1006 stores a set of instructions executable by the processor 1002, including:
  • An operating system 1016 including a program for processing various basic system services and for performing hardware related tasks
  • the application 1018 includes various programs for implementing the image enhancement method described above, and the program can implement the processing flow in the above embodiments, such as the image enhancement device 800 shown in FIG. 8 or the image enhancement shown in FIG. Device 900.
  • the image enhancement device can acquire a third processed image having a high degree of smoothness and enhanced detail features.
  • the image enhancement device can improve the image noise reduction degree and the enhanced detail feature, thereby making the picture more clear, the overall brightness more uniform, the layering feeling stronger, the contrast higher, and the color more vivid.
  • each of the embodiments of the present application can be implemented by a data processing program executed by a data processing device such as a computer.
  • a data processing program constitutes the present application.
  • a data processing program usually stored in a storage medium is executed by directly reading a program out of a storage medium or by installing or copying the program to a storage device (such as a hard disk and or a memory) of the data processing device. Therefore, such a storage medium also constitutes the present invention.
  • the storage medium can use any type of recording method, such as paper storage medium (such as paper tape, etc.), magnetic storage medium (such as floppy disk, hard disk, flash memory, etc.), optical storage medium (such as CD-ROM, etc.), magneto-optical storage medium ( Such as MO, etc.).
  • the present application therefore also discloses a non-volatile storage medium in which is stored a data processing program for performing any or part of the above-described application image enhancement methods of the present application.
  • the method steps described in this application can be implemented by a data processing program, and can also be implemented by hardware, for example, by logic gates, switches, application specific integrated circuits (ASICs), programmable logic controllers, and embedded control. And so on. Therefore, such hardware that can implement the image enhancement method described herein can also constitute the present application.

Landscapes

  • Physics & Mathematics (AREA)
  • General Physics & Mathematics (AREA)
  • Engineering & Computer Science (AREA)
  • Theoretical Computer Science (AREA)
  • Image Processing (AREA)

Abstract

本申请公开了图像增强方法、数据处理设备及存储介质。其中，图像增强方法，由数据处理设备执行，包括：对原始图像进行保边滤波处理而得到第一处理图像；获取原始图像的细节特征；根据所述细节特征和所述第一处理图像，确定第二处理图像；将所述原始图像作为导向图，基于导向图滤波方式对所述第二处理图像进行处理而得到第三处理图像。

Description

图像增强方法、数据处理设备及存储介质
本申请要求于2018年02月13日提交中国专利局、申请号为201810149722.1、发明名称为“图像增强方法、装置、计算设备及存储介质”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
本申请涉及图像处理技术领域，尤其涉及图像增强方法、数据处理设备及存储介质。
背景技术
随着多媒体技术的发展，各个领域对视频有大量需求。为了提供高质量的视频，一些应用场景通常针对视频中图像帧序列进行图像增强。例如，一些图像处理应用可以采用平滑滤波等方式进行图像增强。然而，现有的图像增强方式的增强效果有待提高。
发明内容
本申请提出了一种图像增强方案，以提高图像增强效果。
根据本申请一个方面，提供一种图像增强方法，包括：对原始图像进行保边滤波处理而得到第一处理图像；获取所述原始图像的细节特征；根据所述细节特征和所述第一处理图像确定第二处理图像；将所述原始图像作为导向图，基于导向图滤波方式对所述第二处理图像进行处理而得到第三处理图像。
根据本申请另一个方面，提供一种图像增强装置，包括：第一平滑单元，用于对原始图像进行保边滤波处理而得到第一处理图像；细节获取单元，用于获取所述原始图像的细节特征；图像叠加单元，用于根据所述细节特征和所述第一处理图像确定第二处理图像；图像增强单元，用于将所述原始图像作为导向图，基于导向图滤波方式对所述第二处理图像进行处理而得到第三处理图像。
根据本申请另一个方面，提供一种视频服务系统，包括：根据本申请的图像增强装置。
根据本申请另一个方面，提供一种终端设备，包括根据本申请的图像增强装置。根据本申请另一个方面，提供一种数据处理设备，包括：一个或多个处理器、存储器以及一个或多个程序。一个或多个程序存储在该存储器中并被配置为由所述一个或多个处理器执行，所述一个或多个程序包括用于执行本申请的图像增强方法的指令。
根据本申请另一个方面，提供一种存储介质，存储有一个或多个程序，所述一个或多个程序包括指令，所述指令当由数据处理设备执行时，使得所述数据处理设备执行本申请的图像增强方法。
综上，根据本申请的图像增强方案可以获取具有高平滑程度和增强的细节特征的第三处理图像。这里，图像增强方案通过提高图像的降噪程度和增强细节特征，从而可以使得画面更加清晰、整体亮度更均匀、层次感更强、对比度更高和色彩更鲜艳等效果。
附图说明
为了更清楚地说明本申请实施例中的技术方案，下面将对实施例描述中所需要使用的附图作简单地介绍，显而易见地，下面描述中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性的前提下，还可以根据这些附图获得其他的附图。
图1A示出了根据本申请一些实施例的应用场景的示意图；
图1B示出了根据本申请一些实施例的应用场景的示意图；
图2A示出了根据本申请一些实施例的图像增强方法200的流程图；
图2B和图2C分别示出了根据本申请一些实施例的原始图像和第一处理图像；
图2D示出了根据本申请一些实施例的第二处理图像；
图2E示出了根据本申请一些实施例的第三处理图像；
图2F示出图2E中区域Z2与图2B中Z1的对比图；
图3示出了根据本申请一些实施例的获取第三处理图像的方法300的流程图；
图4示出了根据本申请一些实施例的确定第一线性关系的斜率的方法400的流程图；
图5示出了根据本申请一些实施例的确定第一线性关系的斜率的方法500的流程图；
图6示出了根据本申请一些实施例的确定权重值的方法600的流程图；
图7示出了根据本申请一些实施例的图像增强方法700的流程图；
图8示出了根据本申请一些实施例的图像增强装置800的示意图；
图9示出了根据本申请一些实施例的图像增强装置900的示意图；以及
图10示出了一个数据处理设备的组成结构图。
具体实施方式
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，显然，所描述的实施例仅是本申请一部分实施例，而不是全部的实施例。基于本申请中的实施例，本领域普通技术人员在没有做出创造性劳动前提下所获得的所有其他实施例，都属于本申请保护的范围。
图1A示出了根据本申请一些实施例的应用场景的示意图。如图1A所示，应用场景包括终端设备110和视频服务系统120。这里，终端设备110例如可以是台式计算机、移动电话、平板电脑、智能电视或膝上计算机等各种设备。终端设备110可以包括视频应用111。比如，终端设备110可以下载安装视频应用111。视频应用111也可以预装在终端设备110上。视频服务系统120例如可以包括一个或多个服务器。在一些实施例中，视频服务系统120可以存储视频内容。视频应用111可以通过网络130从视频服务系统120获取视频内容。视频应用111可以对来自本地或来自视频服务系统120的视频内容进行视频增强处理。视频应用111例如可以是视频播放客户端(例如腾讯视频客户端或爱奇艺 视频客户端等)、社交应用或浏览器等各种包含视频播放功能的应用，本申请对此不做限制。
图1B示出了根据本申请一些实施例的应用场景的示意图。如图1B所示，应用场景包括终端设备140和视频服务系统150。视频服务系统150可以包括视频处理应用151。视频处理应用151可以对视频进行图像增强处理。终端设备140可以包括视频播放应用141。视频播放应用141例如是视频播放客户端、社交应用或者浏览器等各种包含视频播放功能的应用。视频播放应用141可以从视频服务系统150获取经过图像增强的视频并进行播放。
综上，终端设备110和视频服务系统150均可以用于对视频进行增强处理。另外需要说明的是，视频增强处理的各种设备可以对视频中各图像帧进行图像增强处理。下面结合图2对根据本申请的图像增强方式进行详细说明。
图2A示出了根据本申请一些实施例的图像增强方法200的流程图。图像增强方法200例如由终端设备110或者视频服务系统150等数据处理设备执行。如图2A所示，图像增强方法200可以包括步骤S201，对原始图像进行保边滤波(Edge-preserving Filtering)处理而得到第一处理图像。在一个实施例中，步骤S201将原始图像作为导向图，基于导向图滤波(Guided Image Filtering)方式对原始图像进行处理而得到第一处理图像。在一些实施例中，步骤S201可以采用中值滤波(The median filtering)、双边滤波(Bilateral Filtering)或者加权最小二乘法滤波(Weighted Least Square Filtering)等各种保边滤波方式进行处理。通过保边滤波处理，步骤S201可以对原始图像进行去噪处理。例如，图2B和图2C分别示出了一些实施例中的原始图像和第一处理图像。如图2B和图2C所示，第一处理图像比原始图像有更高的平滑度。
在步骤S202中，获取原始图像的细节特征。这里，细节特征例如可以包括边缘特征和纹理特征等。在一些实施例中，步骤S202可以首先确定原始图像与第一处理图像之间的颜色差异，然后根据颜色差异确定细节特征。在一些实施例中，步骤S202可以确定原始图像与第一处理图像在一个或多个颜色通道上的 差值，并将所述一个或多个颜色通道上的差值作为颜色差异。
在一些实施例中，在原始图像为灰度图时，步骤S202可以确定原始图像与第一处理图像的灰度值差异，并将其作为颜色差异。
在一些实施例中，原始图像可以包括多个颜色通道。步骤S202可以确定原始图像在多个颜色通道上与第一处理图像的差值，并将差值作为颜色差异。例如，原始图像为红绿蓝(RGB)格式。相应地，步骤S202所获取的颜色差异可以包括红色通道的差值、绿色通道的差值和蓝色通道的差值。这里，步骤S202所确定的颜色差异例如可以包括原始图像中边缘和纹理等特征信息。
在确定颜色差异之后，步骤S202可以根据颜色差异确定原始图像的细节特征。例如，步骤S202可以将颜色差异直接作为细节特征。又例如，步骤S202可以按照调节系数增强颜色差异，以获取增强的颜色差异，并将增强的颜色差异作为原始图像对应的细节特征。这里，调节系数值越大，增强程度越高。另外说明的是，步骤S202也可以采用其他合适的方式获取原始图像的细节特征。
在步骤S202确定细节特征之后，图像增强方法200可以执行步骤S203。在步骤S203中，根据细节特征和第一处理图像，确定第二处理图像。例如，步骤S203可以将细节特征叠加到第一处理图像，而得到第二处理图像。另外，步骤S203也可以采用其他方式来利用细节特征和第一处理图像确定第二处理图像。这里，通过利用细节特征和第一处理图像确定第二处理图像，步骤S203既可以使得第二处理图像保持与第一处理图像接近的平滑程度(即降噪程度)，也可以使得第二处理图像具有增强的细节特征。换言之，第二处理图像中细节更容易被区分。例如,图2D示出了图2C中第一处理图像经过步骤S202的处理而得到的第二处理图像。与图2C相比，图2D中第二处理图像细节更容易被区分。
在步骤S204中，将原始图像作为导向图，基于导向图滤波方式对第二处理图像进行处理而得到第三处理图像。这样，步骤S204可以对第二处理图像进行降噪，以便提高第三处理图像的降噪程度。另外，通过将原始图像作为导向图，步骤S204在第二处理图像进行平滑处理时，可以按照原始图像的细节特征模式 而保留第二处理图像中细节特征。换言之，步骤S204通过利用原始图像作为导向图而对第二处理图像进行图像增强处理，可以使得第三处理图像中细节特征与原始图像的整体概貌更加一致。简言之，步骤S204可以使得第三处理图像中细节特征更加趋向于原始图像的整体概貌，即细节与整体图像更加融合。例如图2E示出了图2D中第二处理图像对应的第三处理图像。图2F示出图2E中区域Z2与图2B中Z1的对比图。区域Z1和区域Z2的内容相同。与代表原始图像的Z1相比，代表第三处理图像的Z2画面更加清晰、整体亮度更均匀、层次感更强、对比度更高。应注意，图2D-2F被显示为灰度图，实际上也可以是彩色图像。步骤S204还可以使得彩色的第三处理图像的色彩更鲜艳。
综上，图像增强方法200通过上述步骤的组合，可以获取具有高平滑程度和增强的细节特征的第三处理图像。这里，图像增强方法200通过提高图像的降噪程度和增强细节特征，从而可以使得画面更加清晰、整体亮度更均匀、层次感更强、对比度更高和色彩更鲜艳等效果。其中，对比度越高，图像的色阶越高，画面越饱满。特别说明的是，图像增强方法200通过上述步骤的组合，可以实现对图像内容的自适应处理，从而提高了图像处理的鲁棒性。另外，图像增强方法200通过采用导向图滤波方式，可以使得第三处理图像的亮度更加均匀。
另外说明的是，上述步骤S201和步骤S204中涉及的导向图滤波方式均可以是：在各颜色通道(例如RGB)上分别进行滤波处理。滤波处理的结果为各颜色通道处理结果的叠加结果。这样，图像增强方法200可以避免细节增强后图像出现色偏。下面以步骤S204的处理过程为例进一步对导向图滤波进行说明。图3示出了根据本申请一些实施例的获取第三处理图像的方法300的流程图。
首先说明的是，为了更好区分本实施例的导向图(即原始图像)、第二处理图像和第三处理图像中像素点，本实施例将导向图中像素点称为第一像素点，将第二处理图像中像素点称为第二像素点，以及将第三处理图像中像素点称为第三像素点。
如图3所示，获取第三处理图像的方法300可以包括步骤S301，对于导向图中每一个第一像素点，确定包含该第一像素点的各第一窗口与第三处理图像中相应位置的第二窗口之间的第一线性关系。这里，任一个第一窗口对应的第一线性关系用于描述该第一窗口中各第一像素点与相应的第二窗口中相应位置像素点的线性关系。其中，各第一窗口的尺寸与相应位置的第二窗口的尺寸均为第一尺寸。这里，第一尺寸用于描述窗口的边长。第一尺寸也可以用半径表示。例如，半径为r，第一尺寸可以表示为(2r+1)。相应地，满足第一尺寸的窗口可以包含(2r+1)*(2r+1)个像素点。
在一些实施例中，步骤S301所确定的第一线性关系可以表示为下述公式：
Figure PCTCN2019074067-appb-000001
其中，w k表示导向图中以像素点k为中心的第一窗口，I i表示第一窗口中第i个第一像素点，q i表示第三处理图像中与I i位置对应的一个第三像素点，a k表示第一线性关系的斜率，b k表示截距。
在一些实施例中，步骤S301可以通过方法400确定第一线性关系的斜率。图4示出了根据本申请一些实施例的确定第一线性关系的斜率的方法400的流程图。
在步骤S401中，确定第一窗口对应的方差。
在步骤S402中，确定第一窗口与第二处理图像中相应位置的第三窗口的协方差。
在步骤S403中，确定方差与正则化系数阈值之和。其中，正则化系数阈值为一个非常小的正数。
在步骤S404中将步骤S402得到的协方差与步骤S403得到的和的比值作为斜率。
在一些实施例中，步骤S301可以根据下述公式确定第一线性关系的斜率：
Figure PCTCN2019074067-appb-000002
其中，a k表示斜率，w k表示以像素点k为中心的第一窗口，I i表示第一窗口中第i个像素点的颜色值，p i表示第三窗口中第i个像素点的颜色值，第三窗口为第二处理图像中与第一窗口位置对应的窗口，u k表示第一窗口中所有像素点的颜色均值，
Figure PCTCN2019074067-appb-000003
表示第三窗口中所有像素点的颜色均值，
Figure PCTCN2019074067-appb-000004
表示第一窗口中所有像素点的颜色值方差，ε表示正则化系数阈值，|w|表示第一窗口中像素点总数。
在一些实施例中，步骤S301可以通过方法500确定第一线性关系的斜率。图5示出了根据本申请一些实施例的确定第一线性关系的斜率的方法500的流程图。
如图5所示，在步骤S501中，确定第一窗口对应的方差。这里，第一窗口对应的方差是指第一窗口中所有像素点的颜色信息的方差。
在步骤S502中，确定第一窗口与第二处理图像中相应位置的第三窗口的协方差。
在步骤S503中，确定第一窗口的细节特征在第一处理图像中的权重值。在一些实施例中，步骤S503可以实施为方法600。图6示出了根据本申请一些实施例的确定权重值的方法600的流程图。在步骤S601中，确定导向图中各第一像素点所对应的第一方差和所对应的第二方差。其中，一个第一像素点对应的第一方差表示以该第一像素点为中心且尺寸满足第二尺寸的窗口对应的方差。一个第一像素点对应的第二方差表示以该第一像素点为中心且尺寸满足所述第一尺寸的窗口对应的方差。这里，第二尺寸例如可以表示图像处理过程所使用的最小窗口尺寸，例如为3，即半径为1，但不限于此。在步骤S602中，确定第一窗口的中心点所对应的第一方差和所对应的第二方差之积，并将该积的二 分一次方作为第三值。在步骤S603中，确定导向图中各第一像素点所对应的第一方差和所对应的第二方差之积，并将各第一像素点对应的所述积的二分之一次方的平均值作为第四值。在步骤S604中，将第三值与第四值的比值作为权重值。
在步骤S504中，确定正则化系数阈值与权重值的第一比值。
在步骤S505中，将步骤S502中协方差与步骤S504中第一比值之和作为第一值，将步骤S502中方差与步骤S504中第一比值之和作为第二值，并将第一值与第二值的比值作为所要确定的斜率。需要说明的是，方法500在确定权重值的过程中，通过使用第二方差(即，满足第二尺寸的窗口内颜色值的方差)确定权重值且利用权重值确定斜率。这样，图像增强方法200在利用与该斜率有关的第二线性关系获取第三处理图像时，可以使得第三处理图像始终保留原始图像中最强细节特征。这里，细节梯度值越大，细节强度越高。换言之，细节特征可以按照细节强度进行区分。图像增强方法200可以始终保留原始图像中最强的细节特征。
在另一些实施例中，步骤S301可以根据下述公式确定第一线性关系的斜率：
Figure PCTCN2019074067-appb-000005
其中，a k表示斜率，w k表示以像素点k为中心的第一窗口，I i表示第一窗口中第i个像素点的颜色值，p i表示第三窗口中第i个像素点的颜色值，第三窗口为第二处理图像中与所述第一窗口位置对应的窗口，u k表示第一窗口中所有像素点的颜色均值，
Figure PCTCN2019074067-appb-000006
表示第三窗口中所有像素点的颜色均值，
Figure PCTCN2019074067-appb-000007
表示第一窗口中所有像素点的颜色值方差，ε表示正则化系数阈值，|w|表示第一窗口中像素点总数，weight k表示第一窗口的权重值。
其中，步骤S301可以根据下述公式确定第一窗口的权重值：
Figure PCTCN2019074067-appb-000008
其中，n表示第一窗口的窗口半径，σ k|r＝n 2表示第一窗口中所有像素点的颜色值方差，σ k|r＝1 2表示以像素点k为中心且半径为1的窗口中所有像素点的颜色值方差，σ i|r＝1 2表示以第一窗口中第i个像素点为中心且半径为1的窗口中所有像素点的颜色值方差，σ i|r＝n 2表示以像素点k为中心且半径为1的窗口中所有像素点的颜色值方差，I表示导向图，|w I|表示导向图中像素点总数。
步骤S301在确定包含一个第一像素点的各第一窗口对应的第一线性关系的斜率后，还可以根据该斜率确定该第一线性关系的截距。在一些实施例中，步骤S301可以根据下述公式确定截距：
Figure PCTCN2019074067-appb-000009
其中，
Figure PCTCN2019074067-appb-000010
表示第二处理图像中与第一窗口位置相对的第三窗口中像素点的颜色均值，a k表示斜率，u k表示第一窗口中所有像素点的颜色均值。
在步骤S302中，根据各第一窗口对应的第一线性关系，确定该第一像素点对应的第二线性关系。其中，第二线性关系用于描述该第一像素点与第三处理图像中相应位置像素点的线性关系。这里，第二线性关系的斜率为各第一窗口对应的第一线性关系的斜率的均值。
在步骤S303中，根据导向图中各第一像素点的颜色值和各第一像素点所对应的第二线性关系，确定第三处理图像中各像素点的颜色值，即获取第三处理图像。在一些实施例中，步骤S303可以根据下述公式确定各像素点的颜色值：
Figure PCTCN2019074067-appb-000011
其中，
Figure PCTCN2019074067-appb-000012
其中，a k表示斜率，w k表示导向图中以像素点k为中心的第一窗口，I i表示第一窗口中第i个像素点的颜色值，|w|表示第一窗口中像素点总数，b k表示一个第一窗口对应的第一线性关系的截距，q i表示第三处理图像中与第一窗口位置对应的第二窗口中第i个像素点的颜色值，
Figure PCTCN2019074067-appb-000013
表示包含像素点i的各第一窗口的斜率的均值，即第二线性关系的斜率，
Figure PCTCN2019074067-appb-000014
表示包含像素点i的各第一窗口的截距的均值，即第二线性关系的截距。
图7示出了根据本申请一些实施例的图像增强方法700的流程图。图像增强方法700例如可以由终端设备110或者视频服务系统150等数据处理设备执行。
如图7所示，图像增强方法700可以包括步骤S701至S704。步骤S701至S704的实施方式分别与步骤S201-S204一致，这里不再赘述。
另外，图像增强方法700还可以包括步骤S705，对第三处理图像进行保边滤波处理而得到第四处理图像。在一些实施例中，步骤S705将第三处理图像作为导向图，基于导向图滤波方式对第三处理图像进行处理而得到第四处理图像。这样，步骤S705通过将第三处理图像自身作为导向图进行保边滤波处理，可以进一步提高降噪程度，从而提高图像增强的效果。
图8示出了根据本申请一些实施例的图像增强装置800的示意图。终端设备110或者视频服务系统150等数据处理设备可以包括图像增强装置800。
如图8所示，图像增强装置800可以包括第一平滑单元801、细节获取单元802、图像叠加单元803和图像增强单元804。
第一平滑单元801用于对原始图像进行保边滤波处理而得到第一处理图像。在一些实施例中，第一平滑单元801将原始图像作为导向图，基于导向图滤波方式对原始图像进行处理而得到第一处理图像。
细节获取单元802用于获取原始图像的细节特征。
在一些实施例中，细节获取单元802可以确定原始图像与第一处理图像之间的颜色差异，根据颜色差异确定细节特征。在一些实施例中，细节获取单元802可以确定原始图像与第一处理图像在一个或多个颜色通道上的差值，并将差值作为颜色差异。在一些实施例中，原始图像为灰度图像。细节获取单元802可以确定原始图像与第一处理图像的灰度值差异，并将其作为颜色差异。在另一些实施例中，细节获取单元802可以按照调节系数增强颜色差异，以获取增强的颜色差异，并将增强的颜色差异作为细节特征。在另一些实施例中，细节获取单元802可以将颜色差异作为细节特征。
图像叠加单元803用于根据细节特征和第一处理图像，确定第二处理图像。
图像增强单元804将原始图像作为导向图，基于导向图滤波方式对第二处理图像进行处理而得到第三处理图像。
在一些实施例中，对于导向图中每一第一像素点，图像增强单元804可以确定包含该第一像素点的各第一窗口与所述第三处理图像中相应位置的第二窗口之间的第一线性关系。其中，各第一窗口的尺寸与相应位置的第二窗口的尺寸均为第一尺寸。
在一些实施例中，图像增强单元804可以确定第一窗口对应的方差。在此基础上，图像增强单元804可以确定方差与正则化系数阈值之和。另外，图像增强单元804还可以确定第一窗口与第二处理图像中相应位置的第三窗口的协方差。这样，图像增强单元804可以将协方差与上述和的比值作为第一窗口对应的第一线性关系的斜率。在一些实施例中，图像增强单元804可以根据下述公式确定第一线性关系的斜率：
Figure PCTCN2019074067-appb-000015
其中，a k表示所述斜率，w k表示以像素点k为中心的所述第一窗口，I i表 示所述第一窗口中第i个像素点的颜色值，p i表示所述第三窗口中第i个像素点的颜色值，所述第三窗口为所述第二处理图像中与所述第一窗口位置对应的窗口，u k表示所述第一窗口中所有像素点的颜色均值，
Figure PCTCN2019074067-appb-000016
表示所述第三窗口中所有像素点的颜色均值，
Figure PCTCN2019074067-appb-000017
表示所述第一窗口中所有像素点的颜色值方差，ε表示正则化系数阈值，|w|表示所述第一窗口中像素点总数。另外，图像增强单元804也可以根据斜率确定第一线性关系的截距。
在另一些实施例中，图像增强单元804可以确定第一窗口对应的方差。图像增强单元804可以确定第一窗口与第二处理图像中相应位置的第二窗口的协方差。另外，图像增强单元804可以确定第一窗口的细节特征在第一处理图像中的权重值。
在一些实施例中，图像增强单元804可以确定导向图中各第一像素点所对应的第一方差和所对应的第二方差。其中，第一方差表示以该第一像素点为中心且尺寸满足第二尺寸的窗口的方差。第二方差表示以该第一像素点为中心且尺寸满足第一尺寸的窗口的方差。在此基础上，图像增强单元804可以确定第一窗口的中心点所对应的第一方差和所对应的第二方差之积，并将积的二分一次方作为第三值。另外，图像增强单元804可以确定导向图中各第一像素点所对应的第一方差和所对应的第二方差之积，并将各第一像素点对应的积的二分之一次方的平均值作为第四值。这样，图像增强单元804将第三值与第四值的比值作为权重值。
在上述基础上，图像增强单元804可以确定正则化系数阈值与权重值的第一比值。图像增强单元804可以将协方差与第一比值之和作为第一值，将方差与第一比值之和作为第二值。这样，图像增强单元804可以将第一值与第二值的比值作为第一窗口对应的第一线性关系的斜率。
在一些实施例中，图像增强单元804可以根据下述公式确定第一线性关系的斜率：
Figure PCTCN2019074067-appb-000018
其中，a k表示所述斜率，w k表示以像素点k为中心的所述第一窗口，I i表示所述第一窗口中第i个像素点的颜色值，p i表示第三窗口中第i个像素点的颜色值，所述第三窗口为所述第二处理图像中与所述第一窗口位置对应的窗口，u k表示所述第一窗口中所有像素点的颜色均值，
Figure PCTCN2019074067-appb-000019
表示所述第三窗口中所有像素点的颜色均值，
Figure PCTCN2019074067-appb-000020
表示所述第一窗口中所有像素点的颜色值方差，ε表示正则化系数阈值，|w|表示所述第一窗口中像素点总数，所述weight k表示所述第一窗口的权重值。
其中，图像增强单元804可以根据下述公式确定第一窗口的权重值：
Figure PCTCN2019074067-appb-000021
其中，所述n表示所述第一窗口的窗口半径，σ k|r＝n 2表示所述第一窗口中所有像素点的颜色值方差，σ k|r＝1 2表示以所述像素点k为中心且半径为1的窗口中所有像素点的颜色值方差，σ i|r＝1 2表示以所述第一窗口中第i个像素点为中心且半径为1的窗口中所有像素点的颜色值方差，σ i|r＝n 2表示以所述像素点k为中心且半径为1的窗口中所有像素点的颜色值方差，I表示所述导向图，|w I|表示所述导向图中像素点总数。
在确定各第一窗口对应的所述第一线性关系之后，图像增强单元804可以根据各第一窗口对应的第一线性关系，而确定第一像素点对应的第二线性关系。其中，第二线性关系用于描述该第一像素点与第三处理图像中相应位置像素点的线性关系。
在上述基础上，图像增强单元804可以根据导向图中各第一像素点的颜色值和各第一像素点所对应的第二线性关系，确定第三处理图像中各像素点的颜色值。综上，图像增强装置800可以获取具有高平滑程度和增强的细节特征的第三处理图像。这里，图像增强装置800通过提高图像的降噪程度和增强的细节特征，从而可以使得画面更加清晰、整体亮度更均匀、层次感更强、对比度更高和色彩更鲜艳等效果。其中，对比度越高，图像的色阶越高，画面越饱满。特别说明的是，图像增强装置800可以实现对图像内容的自适应处理，从而提高了图像处理的鲁棒性。另外，图像增强装置800通过采用导向图滤波方式，可以使得第三处理图像的亮度更加均匀。
图9示出了根据本申请一些实施例的图像增强装置900的示意图。终端设备110或者视频服务系统150等数据处理设备可以包括图像增强装置900。
如图9所示，图像增强装置900可以包括第一平滑单元901、细节获取单元902、图像叠加单元903和图像增强单元904。这里，第一平滑单元901、细节获取单元902、图像叠加单元903和图像增强单元904的实施方式分别与第一平滑单元801、细节获取单元802、图像叠加单元803和图像增强单元804一致，这里不再赘述。
另外，图像增强装置900还可以包括第二平滑单元905，用于对所第三处理图像进行保边滤波处理而得到第四处理图像。在一些实施例中，第二平滑单元905将第三处理图像作为导向图，基于导向图滤波方式对第三处理图像进行处理而得到第四处理图像。这样，通过将第三处理图像自身作为导向图进行保边滤波处理，图像增强装置900可以进一步提高降噪程度，从而提高图像增强的效果。
图10示出了一个数据处理设备的组成结构图。数据处理设备可以实现为终端设备110或者视频服务系统150。如图10所示，数据处理设备包括一个或者多个处理器(CPU)1002、通信模块1004、存储器1006、用户接口1010，以及用于互联这些组件的通信总线1008。
处理器1002可通过通信模块1004接收和发送数据以实现网络通信和/或本地通信。
用户接口1010包括一个或多个输出设备1012，其包括一个或多个扬声器和/或一个或多个可视化显示器。用户接口1010也包括一个或多个输入设备1014。用户接口1010例如可以接收遥控器的指令，但不限于此。
存储器1006可以是高速随机存取存储器，诸如DRAM、SRAM、DDR RAM、或其他随机存取固态存储设备；或者非易失性存储器，诸如一个或多个磁盘存储设备、光盘存储设备、闪存设备，或其他非易失性固态存储设备。
存储器1006存储处理器1002可执行的指令集，包括：
操作系统1016，包括用于处理各种基本系统服务和用于执行硬件相关任务的程序；
应用1018，包括用于实现上述图像增强方法的各种程序，这种程序能够实现上述各实施例中的处理流程，比如可以包括图8所示的图像增强装置800或图9所示的图像增强装置900。这样，图像增强装置可以获取具有高平滑程度和增强的细节特征的第三处理图像。这里，图像增强装置通过提高图像的降噪程度和增强的细节特征，从而可以使得画面更加清晰、整体亮度更均匀、层次感更强、对比度更高和色彩更鲜艳等效果。
另外，本申请的每一个实施例可以通过由数据处理设备如计算机执行的数据处理程序来实现。显然，数据处理程序构成了本申请。
此外，通常存储在一个存储介质中的数据处理程序通过直接将程序读取出存储介质或者通过将程序安装或复制到数据处理设备的存储设备(如硬盘和或内存)中执行。因此，这样的存储介质也构成了本发明。存储介质可以使用任何类型的记录方式，例如纸张存储介质(如纸带等)、磁存储介质(如软盘、硬盘、闪存等)、光存储介质(如CD-ROM等)、磁光存储介质(如MO等)等。
因此本申请还公开了一种非易失性存储介质，其中存储有数据处理程序，该数据处理程序用于执行本申请上述应用图像增强方法的任意一种或部分实施 例。
另外，本申请所述的方法步骤除了可以用数据处理程序来实现，还可以由硬件来实现，例如，可以由逻辑门、开关、专用集成电路(ASIC)、可编程逻辑控制器和嵌微控制器等来实现。因此这种可以实现本申请所述图像增强方法的硬件也可以构成本申请。
以上所述仅为本申请的较佳实施例而已，并不用以限制本申请，凡在本申请的精神和原则之内，所做的任何修改、等同替换、改进等，均应包含在本申请保护的范围之内。

Claims (20)

  1. 一种图像增强方法，由数据处理设备执行，所述方法包括：
    对原始图像进行保边滤波处理而得到第一处理图像；
    获取所述原始图像的细节特征；
    根据所述细节特征和所述第一处理图像，确定第二处理图像；以及
    将所述原始图像作为导向图，基于导向图滤波方式对所述第二处理图像进行处理而得到第三处理图像。
  2. 如权利要求1所述的方法，其中，所述方法还包括：对所述第三处理图像进行保边滤波处理而得到第四处理图像。
  3. 如权利要求2所述的方法，其中，所述对所述第三处理图像进行保边滤波处理而得到第四处理图像，包括：
    将所述第三处理图像作为导向图，基于所述导向图滤波方式对第三处理图像进行处理而得到所述第四处理图像。
  4. 如权利要求1所述的方法，其中，所述对原始图像进行保边滤波处理而得到第一处理图像，包括：
    将所述原始图像作为导向图，基于所述导向图滤波方式对所述原始图像进行处理而得到所述第一处理图像。
  5. 如权利要求1所示的方法，其中，所述获取所述原始图像的细节特征，包括：
    确定所述原始图像与所述第一处理图像之间的颜色差异，根据所述颜色差异确定所述细节特征。
  6. 如权利要求5所示的方法，其中，所述确定所述原始图像与所述第一处理图像之间的颜色差异，包括：
    确定所述原始图像与所述第一处理图像在一个或多个颜色通道上的差值，并将所述差值作为所述颜色差异。
  7. 如权利要求6所述的方法，其中，所述原始图像为灰度图像，所述确定所述原始图像与所述第一处理图像在一个或多个颜色通道上的差值，并将所述差值作为所述颜色差异，包括：
    确定所述原始图像与所述第一处理图像的灰度值差异，并将其作为所述颜色差异。
  8. 如权利要求5所述的方法，其中，所述根据所述颜色差异确定所述细节特征，包括：
    按照调节系数增强所述颜色差异，以获取增强的颜色差异，并将所述增强的颜色差异作为所述细节特征。
  9. 如权利要求5所述的方法，其中，所述根据所述颜色差异确定所述细节特征，包括：将所述颜色差异作为所述细节特征。
  10. 如权利要求1所述的方法，其中，所述将所述原始图像作为导向图，基于导向图滤波方式对所述第二处理图像进行处理而得到第三处理图像，包括：
    对于所述导向图中每一个第一像素点，确定包含该第一像素点的各第一窗口与所述第三处理图像中相应位置的第二窗口之间的第一线性关系，其中，各第一窗口的尺寸与相应位置的所述第二窗口的尺寸均为第一尺寸；
    根据所述各第一窗口对应的所述第一线性关系，确定该第一像素点对应的第二线性关系，其中，所述第二线性关系用于描述该第一像素点与所述第三处理图像中相应位置像素点的线性关系；以及
    根据所述导向图中各第一像素点的颜色值和各第一像素点对应的所述第二线性关系，确定所述第三处理图像中各像素点的颜色值。
  11. 如权利要求10所述的方法，其中，所述确定包含该第一像素点的各第一窗口与所述第三处理图像中相应位置的第二窗口之间的第一线性关系，包括：
    对于包含该第一像素点的任一个第一窗口，确定所述第一窗口对应的方差；
    确定所述第一窗口与所述第二处理图像中相应位置的第三窗口的协方差；
    确定所述方差与正则化系数阈值之和；以及
    将所述协方差与所述和的比值作为所述第一窗口对应的所述第一线性关系的斜率。
  12. 如权利要求10所述的方法，其中，所述确定包含该第一像素点的各第一窗口与所述第三处理图像中相应位置的第二窗口之间的第一线性关系，包括：
    确定所述第一窗口对应的方差；
    确定所述第一窗口与所述第二处理图像中相应位置的第二窗口的协方差；
    确定所述第一窗口的细节特征在所述第一处理图像中的权重值；
    确定正则化系数阈值与所述权重值的第一比值；以及
    将所述协方差与所述第一比值之和作为第一值，将所述方差与所述第一比值之和作为第二值，并将所述第一值与所述第二值的比值作为所述第一窗口对应的所述第一线性关系的斜率。
  13. 如权利要求12所述的方法，其中，所述确定所述第一窗口的细节特征在所述第一处理图像中的权重值，包括：
    确定所述导向图中各第一像素点所对应的第一方差和所对应的第二方差，其中，所述第一方差表示以该第一像素点为中心且尺寸满足第二尺寸的窗口的方差，所述第二方差表示以该第一像素点为中心且尺寸满足所述第一尺寸的窗口的方差；
    确定所述第一窗口的中心点所对应的第一方差和所对应的第二方差之积，并将所述积的二分一次方作为第三值；
    确定所述导向图中各第一像素点所对应的第一方差和所对应的第二方差之积，并将所述各第一像素点对应的积的二分之一次方的平均值作为第四值；以及
    将所述第三值与所述第四值的比值作为所述权重值。
  14. 如权利要求10所述的方法，其中，所述确定包含该第一像素点的各第一窗口与所述第三处理图像中相应位置的第二窗口之间的第一线性关系，包括：
    根据下述公式确定所述第一线性关系的斜率：
    Figure PCTCN2019074067-appb-100001
    其中，a k表示所述斜率，w k表示以像素点k为中心的所述第一窗口，I i表示所述第一窗口中第i个像素点的颜色值，p i表示所述第三窗口中第i个像素点的颜色值，所述第三窗口为所述第二处理图像中与所述第一窗口位置对应的窗口，u k表示所述第一窗口中所有像素点的颜色均值，
    Figure PCTCN2019074067-appb-100002
    表示所述第三窗口中所有像素点的颜色均值，
    Figure PCTCN2019074067-appb-100003
    表示所述第一窗口中所有像素点的颜色值方差，ε表示正则化系数阈值，|w|表示所述第一窗口中像素点总数。
  15. 如权利要求10所述的方法，其中，所述确定包含该第一像素点的各第一窗口与所述第三处理图像中相应位置的第二窗口之间的第一线性关系，包括：
    根据下述公式确定所述第一线性关系的斜率：
    Figure PCTCN2019074067-appb-100004
    其中，a k表示所述斜率，w k表示以像素点k为中心的所述第一窗口，I i表示所述第一窗口中第i个像素点的颜色值，p i表示第三窗口中第i个像素点的颜色值，所述第三窗口为所述第二处理图像中与所述第一窗口位置对应的窗口，u k表示所述第一窗口中所有像素点的颜色均值，
    Figure PCTCN2019074067-appb-100005
    表示所述第三窗口中所有像素点的颜色均值，
    Figure PCTCN2019074067-appb-100006
    表示所述第一窗口中所有像素点的颜色值方差，ε表示正则化系数阈值，|w|表示所述第一窗口中像素点总数，所述weight k表示所述第一窗口的权重值；
    其中，根据下述公式确定所述第一窗口的权重值weight k：
    Figure PCTCN2019074067-appb-100007
    其中，所述n表示所述第一窗口的窗口半径，σ k|r＝n 2表示所述第一窗口中所有像素点的颜色值方差，σ k|r＝1 2表示以所述像素点k为中心且半径为1的窗口中所有像素点的颜色值方差，σ i|r＝1 2表示以所述第一窗口中第i个像素点为中心且半径为1的窗口中所有像素点的颜色值方差，σ i|r＝n 2表示以所述像素点k为中心且半径为1的窗口中所有像素点的颜色值方差，I表示所述导向图，|w I|表示所述导向图中像素点总数。
  16. 一种图像增强装置，其特征在于，所述装置包括：
    第一平滑单元，用于对原始图像进行保边滤波处理而得到第一处理图像；
    细节获取单元，用于获取所述原始图像的细节特征；
    图像叠加单元，用于根据所述细节特征和所述第一处理图像确定第二处理图像；以及
    图像增强单元，用于将所述原始图像作为导向图，基于导向图滤波方式对所述第二处理图像进行处理而得到第三处理图像。
  17. 一种视频服务系统，其特征在于，所述系统包括：如权利要求16所述的图像增强装置。
  18. 一种终端设备，其特征在于，所述终端设备包括：如权利要求16所述的图像增强装置。
  19. 一种数据处理设备，包括：
    一个或多个处理器；
    存储器；以及
    一个或多个程序，存储在该存储器中并被配置为由所述一个或多个处理器执行，所述一个或多个程序包括用于执行权利要求1-15中任一项所述的图像增 强方法的指令。
  20. 一种存储介质，存储有一个或多个程序，所述一个或多个程序包括指令，所述指令当由数据处理设备执行时，使得所述数据处理设备执行如权利要求1-15中任一项所述的图像增强方法。
PCT/CN2019/074067 2018-02-13 2019-01-31 图像增强方法、数据处理设备及存储介质 Ceased WO2019157966A1 (zh)

Priority Applications (2)

Application Number Priority Date Filing Date Title
EP19754702.9A EP3754590B1 (en) 2018-02-13 2019-01-31 Image enhancement method, data processing device, and storage medium
US16/902,481 US11263727B2 (en) 2018-02-13 2020-06-16 Image enhancement method, data processing device, and storage medium

Applications Claiming Priority (2)

Application Number Priority Date Filing Date Title
CN201810149722.1A CN108335275B (zh) 2018-02-13 2018-02-13 图像增强方法、装置、计算设备及存储介质
CN201810149722.1 2018-02-13

Related Child Applications (1)

Application Number Title Priority Date Filing Date
US16/902,481 Continuation US11263727B2 (en) 2018-02-13 2020-06-16 Image enhancement method, data processing device, and storage medium

Publications (1)

Publication Number Publication Date
WO2019157966A1 true WO2019157966A1 (zh) 2019-08-22

Family

ID=62929638

Family Applications (1)

Application Number Title Priority Date Filing Date
PCT/CN2019/074067 Ceased WO2019157966A1 (zh) 2018-02-13 2019-01-31 图像增强方法、数据处理设备及存储介质

Country Status (4)

Country Link
US (1) US11263727B2 (zh)
EP (1) EP3754590B1 (zh)
CN (1) CN108335275B (zh)
WO (1) WO2019157966A1 (zh)

Cited By (1)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN115797238A (zh) * 2022-11-08 2023-03-14 北京极感科技有限公司 图像处理方法

Families Citing this family (3)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN108335275B (zh) * 2018-02-13 2019-12-13 腾讯科技（深圳）有限公司 图像增强方法、装置、计算设备及存储介质
CN111899197B (zh) * 2020-08-05 2024-04-30 广州市百果园信息技术有限公司 一种图像增亮去噪方法、装置、移动终端和存储介质
CN115937054A (zh) * 2022-12-12 2023-04-07 青岛海信日立空调系统有限公司 一种空调系统及其控制方法

Citations (4)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN103337061A (zh) * 2013-07-18 2013-10-02 厦门大学 一种基于多次引导滤波的图像去雨雪方法
KR20130126026A (ko) * 2012-05-10 2013-11-20 중앙대학교 산학협력단 세부정보 추정에 의한 영상 개선 장치 및 방법
CN104318527A (zh) * 2014-10-21 2015-01-28 浙江工业大学 基于小波变换和引导滤波器的医学超声图像去噪方法
CN108335275A (zh) * 2018-02-13 2018-07-27 腾讯科技（深圳）有限公司 图像增强方法、装置、计算设备及存储介质

Family Cites Families (18)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN102064796B (zh) * 2010-11-15 2014-09-17 中南大学 一种简化的具有负系数的加权重复伪中值滤波方法
US9002131B2 (en) * 2011-09-18 2015-04-07 Forus Health Pvt. Ltd. Method and system for enhancing image quality
KR101401916B1 (ko) 2012-05-02 2014-06-10 민진식 회전 직립 자전거 거치대
KR101736468B1 (ko) * 2012-12-24 2017-05-29 한화테크윈 주식회사 영상 처리 장치 및 방법
EP2976877B1 (en) * 2013-03-22 2020-01-08 Synaptics Incorporated Guided filter-based detail enhancement
US9754356B2 (en) * 2013-04-12 2017-09-05 Agency For Science, Technology And Research Method and system for processing an input image based on a guidance image and weights determined thereform
JP6287100B2 (ja) * 2013-11-20 2018-03-07 株式会社リコー 画像処理装置、画像処理方法、プログラム及び記憶媒体
US9911179B2 (en) * 2014-07-18 2018-03-06 Dolby Laboratories Licensing Corporation Image decontouring in high dynamic range video processing
US9779491B2 (en) * 2014-08-15 2017-10-03 Nikon Corporation Algorithm and device for image processing
US20160241884A1 (en) * 2015-02-13 2016-08-18 Tandent Vision Science, Inc. Selective perceptual masking via scale separation in the spatial and temporal domains for use in data compression with motion compensation
LU92673B1 (en) * 2015-03-05 2016-09-06 Iee Int Electronics & Eng Sa Method and system for real-time noise removal and image enhancement of high-dynamic range images
JP6164238B2 (ja) * 2015-03-18 2017-07-19 カシオ計算機株式会社 診断装置並びに当該診断装置における画像処理方法及びそのプログラム
US20170109867A1 (en) * 2015-10-16 2017-04-20 Motorola Mobility Llc Camera array for performing non-local means image processing over multiple sequential images
JP6635799B2 (ja) * 2016-01-20 2020-01-29 キヤノン株式会社 画像処理装置、画像処理方法及びプログラム
US10223772B2 (en) * 2016-03-22 2019-03-05 Algolux Inc. Method and system for denoising and demosaicing artifact suppression in digital images
US11113800B2 (en) * 2017-01-18 2021-09-07 Nvidia Corporation Filtering image data using a neural network
CN107316279B (zh) * 2017-05-23 2019-11-22 天津大学 基于色调映射和正则化模型的弱光图像增强方法
CN107481203B (zh) * 2017-08-14 2020-05-15 厦门美图之家科技有限公司 一种图像导向滤波方法及计算设备

Patent Citations (4)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
KR20130126026A (ko) * 2012-05-10 2013-11-20 중앙대학교 산학협력단 세부정보 추정에 의한 영상 개선 장치 및 방법
CN103337061A (zh) * 2013-07-18 2013-10-02 厦门大学 一种基于多次引导滤波的图像去雨雪方法
CN104318527A (zh) * 2014-10-21 2015-01-28 浙江工业大学 基于小波变换和引导滤波器的医学超声图像去噪方法
CN108335275A (zh) * 2018-02-13 2018-07-27 腾讯科技（深圳）有限公司 图像增强方法、装置、计算设备及存储介质

Non-Patent Citations (1)

* Cited by examiner, † Cited by third party
Title
See also references of EP3754590A4 *

Cited By (1)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN115797238A (zh) * 2022-11-08 2023-03-14 北京极感科技有限公司 图像处理方法

Also Published As

Publication number Publication date
US20200311880A1 (en) 2020-10-01
EP3754590A1 (en) 2020-12-23
CN108335275B (zh) 2019-12-13
CN108335275A (zh) 2018-07-27
EP3754590A4 (en) 2021-03-24
US11263727B2 (en) 2022-03-01
EP3754590B1 (en) 2022-04-20

Similar Documents

Publication Publication Date Title
US10861133B1 (en) Super-resolution video reconstruction method, device, apparatus and computer-readable storage medium
US12182979B2 (en) Brightness and contrast enhancement for video
US11373275B2 (en) Method for generating high-resolution picture, computer device, and storage medium
US20240121521A1 (en) Image processing based on object categorization
WO2020108009A1 (en) Method, system, and computer-readable medium for improving quality of low-light images
US9672414B2 (en) Enhancement of skin, including faces, in photographs
WO2020093782A1 (en) Method, system, and computer-readable medium for improving quality of low-light images
WO2019157966A1 (zh) 图像增强方法、数据处理设备及存储介质
US10198801B2 (en) Image enhancement using self-examples and external examples
US10096089B2 (en) Accelerated skin smoothing effect
CN106485720A (zh) 图像处理方法和装置
US20250272803A1 (en) Method, computer device, and storage medium for processing video denoising model
CN106462986A (zh) 用于数字图像的颜色处理的方法和系统
US9286653B2 (en) System and method for increasing the bit depth of images
WO2022247702A1 (zh) 图像处理方法及装置、电子设备和存储介质
US12488510B2 (en) Encoders for improved image dithering
US20230127327A1 (en) System and method for learning tone curves for local image enhancement
US8824778B2 (en) Systems and methods for depth map generation
WO2019047409A1 (zh) 图像处理方法、系统、可读存储介质及移动摄像设备
US10992873B2 (en) Systems and methods for color matching for realistic flash images
US12354243B2 (en) Image denoising method, device, and computer-readable medium using U-net
CN114519753B (zh) 图像生成方法、系统、电子设备、存储介质和产品
CN116503263B (zh) 渲染图降噪方法、模型渲染方法、装置、设备及存储介质
CN110166700A (zh) 创建选择性虚拟长曝光图像
CN116917900A (zh) 在像素到像素神经网络中处理数据

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: 19754702

Country of ref document: EP

Kind code of ref document: A1

NENP Non-entry into the national phase

Ref country code: DE

ENP Entry into the national phase

Ref document number: 2019754702

Country of ref document: EP

Effective date: 20200914