JPH01251973A - Orthogonal transform coding device and orthogonal inverse transform decoding device for video signal - Google Patents

Abstract

PURPOSE:To record or transmit picture data with smaller data quantity compared with the data quantity in directly coding frame picture data by compressing the data by obtaining the difference between one field, for example, a second field, and the other field, namely, a first field, when the picture data of the second field are to be coded. CONSTITUTION:A frame video signal outputted from a video signal source 10 is converted into corresponding digital data at an analog-digital converter 12, transferred from a bus 14 to a memory 18, and once stored in the memory 18. The frame picture data are composed of the data of the first field and that of the second field. Here, the data are compressed by obtaining the difference between one field to compose an interlaced-scanned frame, for example, the second field, and the other field, namely, the first field, when the picture data of the second field are to be coded. Thus, the picture data can be recorded or transmitted with the smaller data quantity compared with that in directly coding the frame picture data.

Description

[Detailed Description of the Invention] The present invention relates to an orthogonal transform encoding method for a video signal, in particular, an apparatus for orthogonally transforming and encoding a video signal subjected to field interlaced scanning, and a method for encoding a video signal encoded in this manner. The present invention relates to a device that decodes data and performs orthogonal inverse transformation to form a video signal.

For example, when storing an image signal captured by a solid-state imaging device such as a CCD in a storage device such as a memory card or magnetic disk, the data of the image signal is compressed to a small amount, taking into consideration the capacity of the storage device. It is necessary. 1 like this
Orthogonal transform encoding is known as a method for compressing image data.

This method is as follows. First, an image represented by an image signal is divided into a predetermined number of blocks, and data of each pixel in each divided block is orthogonally transformed.

In the image signal, the low frequency component occupies a large component in terms of power. On the other hand, although high frequency components are not large in terms of power, they are significant in terms of information.

Moreover, the characteristics for these are also visually different.

Therefore, the image signal is orthogonally transformed from such low frequency components to high frequency components, and each component is quantized and encoded. On the receiving or reproducing side, the encoded signal is inversely transformed to obtain the original signal. In this way, efficient encoding can be performed.

In orthogonal transform encoding, the entire screen is divided into a plurality of blocks, with an appropriate number of pixels forming one block, and a numerical string consisting of sample values is orthogonally transformed for each block. In other words, the image that matches the characteristics of the original image signal.
Perform linear transformation with mutually independent transformation axes. As a result, each transformed term becomes more independent, that is, more uncorrelated, than the original sample value. This suppresses redundant information. This method is, so to speak, an operation on the frequency axis.

As a result, power is concentrated in a specific component due to the statistical properties of the image signal. Therefore, while taking visual characteristics into account, many bits are allocated to low-frequency components with high power, and high-frequency components with low power are coarsely quantized using a small number of bits. This allows the number of bits per block to be reduced.

By performing such orthogonal transformation and encoding, the capacity of the storage device can be reduced compared to the case where the pixel data constituting the image signal is quantized and stored as is.

By the way, a field interlaced scanning video signal in which two fields are interlaced scanned to form one frame still has a large degree of redundancy even after such orthogonal transformation and encoding, when viewed in units of l frames. This is because the image data of such a pair of fields is highly correlated,
This is because if this were to be orthogonally transformed and encoded as it is, it would simply require a storage capacity twice as large as the image data storage area for l fields.

Purpose The present invention eliminates the drawbacks of the prior art and provides an orthogonal transform encoding device and an orthogonal inverse transform decoding device for video signals that enable efficient data compression even for field-interlaced video signals. The purpose is to provide.

DISCLOSURE OF THE INVENTION According to the present invention, an interlaced-scanned frame video signal is encoded optimally for the feature that the correlation between the first and second field images it represents is strong. When encoding image data of one field, for example, the second field, data compression is performed by taking the difference from the other field, that is, the first field. This makes it possible to record or transmit image data with a smaller amount of data than when directly encoding frame image data.

The orthogonal transform encoding device for a video signal according to the present invention includes a first storage means for storing image data representing a pair of fields subjected to interlaced scanning, and a screen representing the image data stored in the first storage means. a blocking means that divides the image data into blocks and extracts one image data block by block; an orthogonal transformation means that orthogonally transforms the image data of the extracted block to form transform coefficient data; a difference means that calculates a difference between the fields and forms difference data representing the difference; and an encoding means that encodes the transform coefficient data of one field of the transform coefficient data using a larger number of bits for lower frequency components; and first output means for outputting encoded transform coefficient data and difference data.

Further, the orthogonal inverse transform decoding device for a video signal according to the present invention includes: a third storage means for accumulating data obtained from the first output means; a decoding means for decoding the data in the third storage means block by block; an addition means for adding the decoded transform coefficient data to the difference data, an orthogonal inverse transform of the decoded transform coefficient data for each block to obtain image data of one field, and an orthogonal inverse transform for the added result data for each block; and orthogonal inverse transform means for performing inverse transform to obtain image data of the other field.

DESCRIPTION OF EMBODIMENTS Next, the orthogonal transform encoding method for video signals according to the present invention will be described in detail according to embodiments with reference to the accompanying drawings.

Referring first to FIG. 5, an embodiment of the orthogonal transform coding apparatus for a video signal according to the present invention receives a video signal source 1O, such as a solid-state imaging device, which is converted into an image signal via an analog-to-digital converter (ADC) 12. and is connected to path 14. In this embodiment, the video signal source 10 is a signal source that forms a field-interlaced frame video signal in the form of an analog signal, and is not an imaging device, but a receiving circuit that receives such a signal from a communication line, for example. The analog-to-digital converter 12, which may be a digital converter, is a signal conversion circuit that converts such an analog video signal into corresponding digital data.

This device is an orthogonal transform encoding device that orthogonally transforms and encodes such a frame video signal and stores it in, for example, an image recording memory te.The image recording memory 1B is a memory such as an IC memory or a magnetic disk. The device is advantageously applied. For example, in the case of an IC memory, it may be fixedly mounted inside the device, but it may also be in the form of a card or cartridge that is removably connected to the device via the connector 17.

In this apparatus, an image represented by image data is divided into blocks of a predetermined size, and the data of each pixel in each divided block is orthogonally transformed, quantized, and encoded. At this time, when encoding the image data of one field, for example, the second field, constituting the interlaced scanned frame, the difference between it and the other field, that is, the first field, is calculated. This encoding method is most suitable for the feature that the first and second field images represented by the frame video signal have a strong correlation between them. This makes it possible to record or transmit image data with a smaller amount of data than when directly encoding frame image data.

This data compression function temporarily stores image data obtained from the video signal source 10 in the memory 18, and stores the image data in the memory 18.
This is achieved by calculating and processing the above. The memory 18 constitutes a main storage area that functions as a work area for this purpose, and such processing is configured to be executed by the central processing unit 1l (CAPυ) 22 according to a processing program.

The image data thus compressed in the memory 18 may be stored in the image recording memory 16 or may be transmitted from the transmitter 24 to the line 26.

Both the image recording memory 1B and the transmitter 24 may be provided, or either one may be provided.

The data compression process in the non-embodiment will be explained in more detail with reference to FIG. First, the frame video signal output from the video signal source 10 is sent to the analog-to-digital converter 1.
2, the data is converted into corresponding digital data. This is transferred from bus 14 to memory 1B and stored therein. This frame image data is composed of first and second field data 30a and 30b.

First, the data 30a of the first field is read out, and the block is extracted by the functional unit 32a. More specifically,
The image represented by the image data, even if you think of it as a frame,
It can also be thought of as a field, which is divided into blocks of a predetermined size. In other words, the entire rectangular screen 80 is generally divided into a plurality of square areas, ie, blocks 62a and 82b, as shown in FIG. 7 in data compression processing.
, , , fi2h, etc. Each of these square areas has, for example, 18 pixels in the horizontal direction Y (FIG. 7) and 16 pixels in the vertical direction X of the screen, but it may also be a rectangle with a different shape. , extracts data of each pixel constituting a specific block, for example [i2a, included in the data 30a of the first field, and provides this to the orthogonal transformation function unit 34a.

The state of the pixel data is illustrated in FIG. 8A. The block extraction function unit 32a sequentially performs this block extraction for all blocks in the first field 30a.

The image data 48a of the specific block extracted in this way, for example 82a, is orthogonally transformed by the orthogonal transformation function section 34a. In this embodiment, cosine transformation is advantageously applied to the orthogonal transformation 34a, whereby the image data is transformed into frequency domain data. An example of the orthogonally transformed data is shown in FIG. 8B. Cosine transformation has few transformation errors and is suitable for compressing image data.

Similarly, data 30b of the second field is also stored in the memory 18.
, and the block is extracted by the functional unit 32b. This reading and block extraction 32b are preferably performed in synchronization with those of the first field. If performed synchronously, data compression of the first and second) 4-fields can be performed simultaneously. If synchronization is not performed, reading of data in the first field and block extraction are performed when compressing data in the second field, in addition to data compression processing in the first field.

In any case, the data 30b of the second field is
Similar to the field, block extraction 32b is performed and orthogonal transformation 34b is performed. Block extraction 32b of the second field 30a is performed on the same block as that of the first field 30a, fi2a in this example.

The state of the pixel data of the extracted block 82a of the second field is illustrated in FIG. 8C, which is generally almost the same as that of the first field (FIG. 8B). Thereafter, this block extraction 32b and orthogonal transformation 34b are performed using the first
The second field 3 is synchronized with those of field 30a.
This is performed sequentially for all blocks of 0b. In this way, the image data 30a and 30 of the first and second fields
b is converted into frequency domain data for each block such as 62a, 82b, . . . , e2h, and is held in the memory 18.

The transform coefficient data 40a that has been orthogonally transformed in the first field 34a is first encoded in the encoding function section 38a. This encoding is also performed for each block such as 82a, 62b, . . . , 62h. In this embodiment, the encoding method is to cut off the transform coefficient data to be encoded at an appropriate threshold value, and for the remaining transform coefficient data, add a large number of bits to the low frequency components, and add a large number of bits to the high frequency components, that is, the analog-to-digital converter 12. A method is adopted in which fewer bits are assigned to frequency components closer to the sampling frequency. The encoding function unit 38a quantizes the orthogonal transform data 40a with such encoding bit allocation.

Since the distribution of transform coefficient data generally differs depending on the pattern of the image represented by the video signal, adaptive encoding in which the threshold value is made variable according to the variance of the transform coefficient data is advantageously applied. In other words, in the case of a picture with little gradation change, the frequency of low frequency components is extremely high, so it is preferable to allocate a large number of bits to the low frequency components and set a high threshold value. The opposite is true for detailed patterns.

On the other hand, regarding the second field, the difference 42 from the transform coefficient data 40b transformed by the orthogonal transform 34b with respect to the transform coefficient data 40a of the corresponding block of the first field is extracted in the difference function section 38. In other words, the transform coefficient data 4 of the first field that has been orthogonally transformed 34a
Each frequency component value 64a included in 0a (8th B
The functional unit 38 obtains a difference 42 between the corresponding frequency component value 84b (FIG. 8C) included in the transform coefficient data 40b of the second field which has also been orthogonally transformed. This difference extraction is performed sequentially for all frequency component values included in one block, for example 62a. The state of the difference data is illustrated in FIG. 8D. Thereafter, this operation is performed in the corresponding blocks 62b of both fields.

, , , every 82h, etc.

In this embodiment, this difference data 42 is stored in the encoding function unit 3[
1b. This encoding may also be adaptive.

In this way, the first field data 42a encoded 36a for each block 82a and the like and the differential data 42b also encoded are completed in the memory 18. memory 1
These encoded data 42a and 42b completed in 8
is read out according to instructions from the CPU 22, transferred to and stored in the image recording memory 16, or transmitted to the transmission path 2B by the transmitter 24.

Furthermore, as shown in FIG. 2, in another embodiment of the present invention, the difference data 42 may not be encoded and may be directly output 113/24 together with the encoded data 42a of the first field. In other figures showing other such embodiments of the invention, elements similar to those shown in FIG. 1 are designated by the same reference numerals.

As described above, according to these embodiments, attention is paid to the feature that the interlaced scanned frame video signal has a strong correlation between the images of the first and second fields that it represents. The field is orthogonally transformed and encoded as it is, and the other field, in this example, the second
Data compression is performed for the field by taking the difference from the first field. therefore,
Image data can be recorded or transmitted, that is, outputted, with a smaller amount of data than when frame image data is directly encoded.

As can be seen from FIG. 8D, the difference data subjected to the orthogonal transformation 34b has no DC component in most cases, and the high frequency region is mostly occupied by noise.
Significant values are distributed in the frequency domain representing the difference between the patterns in the field and the second field. This region generally takes the intermediate frequency region. Therefore, in the encoding 313b, a certain number of bits are allocated to a specific intermediate frequency region, and no encoding bits are allocated to other parts, especially the DC component.

Referring to FIG. 3, another embodiment of the present invention is configured to take a difference 44 between the first and second fields after block extraction 32a and 32b and before performing orthogonal transformations 34a and 34b. There is. In this embodiment, each block 82a has gradation data 48a and 46b of the first and second fields 30a and 30'b.
Differences are taken for each.

As in the other embodiments, the first field gradation data 46a is directly subjected to orthogonal transformation 34a and encoding 28a.

However, for the data 30b of the second field, after block extraction 32b is performed, a difference 48 with respect to the gradation data 48a of the first field is taken in a difference function section 44. The orthogonal transformation function section 34b orthogonally transforms this difference data 48, and the encoding section encodes the difference data thus orthogonally transformed. This encoding 36b may also be adaptive encoding, or the encoding 3eb may not be performed.

The image data thus recorded in the image recording memory 1B or sent out to the communication line 26 is reproduced into the original video signal by an orthogonal inverse transform decoding device configured as illustrated in FIG. This orthogonal inverse transform decoding device includes an image recording memory 1
The receiver 70 receives the image data from the communication line 2B, decodes it and performs orthogonal inverse transformation to restore the original video signal, and transmits it to the device to be used. 72 has a function to reproduce this. The hardware configuration may be the same as the orthogonal transform encoding device shown in FIG. 5, that is, the same device may have a recording function and a playback function. It is configured as a separate device from the orthogonal transform encoding device.

The image recording memory 16 in which image data compressed by an orthogonal transform encoding device is recorded is usually.

Although it takes the form of storage devices such as IC memory and magnetic disks,
For example, in the case of an IC memory card or cartridge, if the zero-image recording memory 1B removably connected to the apparatus through the connector 74 is a magnetic disk, the magnetic disk is set in the magnetic disk reading device. Further, when image data is transmitted from the orthogonal transform encoding device through the communication line 16, the image data is transmitted to the receiver 7.
Received as 0.

The data reproducing function of this device is realized by temporarily storing image data obtained from the image recording memory 1B or the communication line 26 in the memory 7B, and performing calculations and processing on the memory 76. The memory 7B serves as a main storage area as a work area for this purpose, and the CPo 80 is configured to perform such processing in accordance with a processing program. The image data reproduced in the memory 7B is output as an analog video signal to the utilization device 72 through a digital-to-analog converter (IIAC) 82 in accordance with instructions from the CPU eo. Note that the image recording memory 16 may have both the connection function, that is, the connector 74 and the receiver 70, or may have only one of them.

The data reproduction process in this embodiment will now be described in more detail with reference to FIG. First, compressed image data 86 read from image memory 18 or received from line 26 at receiver 70, referred to herein as encoded data for convenience, is transferred from bus 84 to memory 76.
is transferred to and stored there.

First, the encoded data 8B is divided into the first field data 88a and the difference data 88b in each block 82a,
82b, . . . , 82h, etc., and decoded by functional units 90a and 90b, respectively. This decoding process is the inverse process of the process performed on the image data by the encoding function units 38a and 38b of the orthogonal transform encoding device, and is performed according to the same decoding rules as those encoding rules. Of course, when reproducing the "encoded" data obtained in the embodiment described with reference to FIG. 2, since the differential data is not "encoded" in the narrow sense, The signal is directly output to the force calculation function section 92 without going through the decoding function section 80b. Here, "encoding in a narrow sense" refers to quantizing by allocating bits appropriate to the nature of image data.

The compressed data of the first field is restored to frequency domain data 9Eia for each block by the decoding function unit 90a,
It is subjected to orthogonal inverse transform 94a. This inverse transformation 94a is sequentially performed for each block 82a, f(2b, .
b, .
and is added to each corresponding frequency component value 84a included in the first field's transform coefficient data 9[ia. Data 98 resulting from the addition is passed to the orthogonal inverse transform function section 94b as transform coefficient data of the second field.

As a result, the converted output e4b of the second field of the block 62a is completed.

The orthogonal inverse transform function section 94a performs orthogonal inverse transform on the decoded transform coefficient data 98a of the first field. In this way, the orthogonal inverse transform 94a converts all blocks 8 of l frame
2a, 62b, . . . , 82h, etc. are sequentially performed. As a result, the tone data 1 of the first field
00a is completed in the memory 76. In other words, the gradation data shown in FIG. 8A is completed. Similarly, the orthogonal inverse transform function section 94
b performs orthogonal inverse transformation on the decoded transform coefficient data 98 of the second field for all blocks 62a.

82b, , , 82b, etc., and then
The data is stored in the memory 76 as field gradation data 100b. The orthogonal inverse transforms 94a and 94b are inverse processing of the orthogonal transform performed on the image data by the orthogonal transform function units 34a and 34b of the orthogonal transform encoding device,
This is done according to the same inverse transformation rules as those transformation rules.

The image data 100a and 100b of the first and second fields completed in the memory 76 are stored in the CP
In response to instructions from U 80, the signals are sequentially read out to digital-to-analog converter 82 using a field interlaced scanning method. The converter 82 converts this into a corresponding analog signal and outputs it to the utilization device 72. The utilization device 72 is, for example, a video playback device including a video monitor device such as a CR-book or an image printer. In this way, the analog video signal is reproduced by the device used. When outputting to a video monitor device,
It is preferable to output the first and second fields using an interlaced scanning method.

Although an embodiment in which the present invention is applied to orthogonal transformation has been described,
The present invention is not limited to this, but can be effectively applied to other encoding methods such as predictive encoding, Hadamard transform, and Fourier transform.

Effects According to the present invention, the image data of one field of an interlaced scanned frame video signal is compressed, and when encoding the image data of the other field, the difference from the one field is calculated. This allows data compression. This makes it possible to record or transmit image data with a smaller amount of data than when directly encoding frame image data.

[Brief explanation of the drawing]

FIG. 1 is a functional block diagram showing the configuration of data compression processing in this embodiment of the orthogonal transform encoding device for video signals according to the present invention, and FIGS. 2 and 3 show other implementations of the orthogonal transform encoding device. FIG. 4 is a functional block diagram showing the configuration of data reproduction processing in the orthogonal inverse transform decoding device for video signals according to the present invention; FIG. 5 is a functional block diagram showing the configuration of the orthogonal transform encoding device for video signals according to the present invention. FIG. 6 is a block diagram showing an embodiment of the orthogonal inverse transform decoding device for video signals according to the present invention; FIG. 7 is an explanatory diagram showing an example of screen block extraction in the embodiment; 8A to 8D are explanatory diagrams showing an example of data compression processing in the same embodiment. Explanation of part symbols +8. ,,image recording memory 18.78. , memory 32a, , block extraction function unit 34a, , orthogonal transformation function unit 36a, , encoding function unit 38.44. , difference function section 90a, ..., decoding function section 92, ..., addition function section 94a, ..., orthogonal inverse transform function section Patent applicant: Fuji Photo Film Co., Ltd. Agent: Takao Katori, Takao Ohiwa, Figure 7Y, Figure 8A

Claims (1)

[Claims] 1. A first storage means for storing image data representing a pair of interlaced scanned fields, and a screen represented by the image data stored in the first storage means is divided into blocks. a blocking means for extracting the image data block by block; an orthogonal transformation means for orthogonally transforming the image data of the extracted block to form transformation coefficient data; and the pair of fields of the transformation coefficient data. a difference means for taking a difference between the two fields and forming difference data representing the difference; and an encoding means for encoding the transform coefficient data of one field of the transform coefficient data using a larger number of bits for lower frequency components. An orthogonal transform encoding device for a video signal, comprising: first output means for outputting the encoded transform coefficient data and the difference data. 2. The apparatus according to claim 1, wherein the encoding means encodes the difference data, and the first output means outputs the encoded difference data. . 3. The apparatus according to claim 1, wherein the output means comprises:
An orthogonal transform encoding device characterized in that it includes a connection means to which a second storage means for accumulating the encoded data is connected. 4. Third storage means for accumulating the data obtained from the first output means of the apparatus according to claim 1, and decoding the transform coefficient data of the data in the third storage means for each block. a decoding means; an addition means for adding the decoded transform coefficient data to the difference data of the data in a third storage means; As the image data of the field,
An orthogonal inverse transform decoding device for a video signal, comprising an orthogonal inverse transform means for orthogonally inverse transforming the data resulting from the addition for each block to obtain image data of the other field. 5. Third storage means for accumulating the encoded data obtained from the first output means of the apparatus according to claim 2; and decoding for decoding the data in the third storage means for each block. means for adding the decoded difference data to transform coefficient data of the decoded data, and performing orthogonal inverse transform on the decoded transform coefficient data for each block to obtain image data of the one field. An orthogonal inverse transform decoding device for a video signal, comprising an orthogonal inverse transform means for performing orthogonal inverse transform on the data resulting from the addition for each block to obtain image data of the other field. 6. The orthogonal inverse transform decoding device according to claim 4 or 5, wherein the first and second orthogonal inverse transform
An orthogonal inverse transform decoding device comprising: second output means for outputting image data of the field as an interlaced scanning video signal.