WO2024098843A1

WO2024098843A1 - Method and system for improving compression quality of jpeg, chip and electronic device

Info

Publication number: WO2024098843A1
Application number: PCT/CN2023/110315
Authority: WO
Inventors: 孙旭; 周玉龙; 刘刚; 李拓
Original assignee: 苏州元脑智能科技有限公司
Priority date: 2022-11-10
Filing date: 2023-07-31
Publication date: 2024-05-16
Also published as: CN115474062B; CN115474062A

Abstract

The present application relates to the field of image processing. Disclosed is a method for improving the compression quality of joint photographic experts group (JPEG), comprising: counting the pixel values in an image to be compressed so as to obtain the pixel value having the largest number, and, according to the pixel value having the largest number, and the total number of pixels and the pixel depth of said image, generating an adjustment coefficient; performing color mode conversion and discrete cosine transform (DCT) on said image so as to obtain a coefficient matrix, wherein before DCT is performed, said image is divided into a plurality of image blocks and adjustment is carried out according to the adjustment coefficient; calculating the brightness expectation corresponding to each adjusted image block, and, on the basis of the brightness expectation, calculating a standard deviation corresponding to each image block; calculating a standard deviation coefficient and a brightness expectation coefficient for each image block; acquiring a standard quantization parameter table specified by the JPEG protocol, and updating the standard quantization parameter table by using the brightness expectation coefficient and the standard deviation coefficient corresponding to each image block; and quantizing said coefficient matrix by using the updated quantization parameter table. Also disclosed in the present application are a system, a chip and an electronic device.

Description

A method, system, chip and electronic device for improving JPEG compression quality

CROSS-REFERENCE TO RELATED APPLICATIONS

This application claims priority to a Chinese patent application filed with the China Patent Office on November 10, 2022, with application number 202211408785.7 and application name “A method, system, chip and electronic device for improving JPEG compression quality”, the entire contents of which are incorporated by reference in this application.

Technical Field

The present application relates to the field of JPEG, and in particular to a method, system, chip and electronic device for improving JPEG compression quality.

Background technique

JPEG (Joint Photographic Experts Group, a compression standard for continuous tone still images) video image compression is a commonly used international standard for video image compression, which is used to compress continuous tone still images (including grayscale images and color images). As shown in Figure 1, the classic JPEG image compression process mainly includes color mode conversion, DCT (Discrete Cosine Transform), quantization, Z-shaped arrangement, run length encoding, Huffman coding, and data packaging.

In the existing technology, the goal is generally to control the JPEG compression bit rate, monitor the actual compressed data bit rate, compare it with the target value, and adjust the quantization coefficient of the entire image based on the difference. However, the starting point of this method is to control the bit rate, and it does not pay attention to the image quality, and in most cases the image quality cannot be optimized. Secondly, although this method can ensure the stability of the bit rate, that is, the output video image is not stuck, but the same set of quantization coefficients is used for the entire image, and the quality of the compressed image varies greatly for different image parts.

Summary of the invention

In view of this, according to a first aspect, an embodiment of the present application proposes a method for improving the quality of Joint Photographic Experts Group JPEG compression, comprising the following steps:

Counting pixel values in the image to be compressed to obtain the largest number of pixel values, and generating an adjustment coefficient according to the number and pixel value corresponding to the largest number of pixel values, the total number of pixels of the image, and the pixel depth;

The image is subjected to color mode conversion and discrete cosine DCT transformation to obtain a coefficient matrix, wherein the image is divided into a plurality of image blocks before the DCT transformation, and the pixel value of each pixel point in each image block is adjusted by adding the adjustment coefficient;

Calculate the brightness expectation corresponding to each image block after adjustment and calculate the standard deviation corresponding to each image block based on the brightness expectation;

Compare the standard deviation with the preset first standard deviation threshold and the second standard deviation threshold to determine the standard deviation coefficient of each image block, and compare the brightness expectation with the preset first brightness expectation threshold and the second brightness expectation threshold to determine the brightness expectation coefficient of each image block;

Obtain a standard quantization parameter table specified in the JPEG protocol and update the standard quantization parameter table using the brightness expectation coefficient and standard deviation coefficient corresponding to each image block;

The coefficient matrix is quantized using the updated quantization parameter table to obtain a quantized coefficient matrix corresponding to each image block.

In the embodiment of the present application, counting the pixel values in the image to be compressed to obtain the largest number of pixel values further includes:

Call the preset dual-port random access memory RAM, where the data width of the dual-port RAM is log2Fream_S, the data depth is 2 ^pix_width , Fream_S is the total number of pixels of the image, and pix_width is the pixel depth;

In response to detecting the beginning of the image frame, all data in the dual-port RAM are set to 0.

In an embodiment of the present application, all data in the dual-port RAM is set to 0, further comprising:

Use reset to set all data in the dual-port RAM to 0.

All data in the dual-port RAM are set to 0 by writing zero to each address in turn.

In the embodiment of the present application, it also includes:

In response to starting to receive pixel value data of an image, assigning the received pixel value pixel_data to the address of port B of the dual-port RAM, and at the same time setting the enable of port B to 1, the read-write control position to 0, performing a read operation, and determining the number of pixel points whose pixel value is pixel_data according to the read result;

Set the enable of port A of the dual-port RAM to 1, the read/write control bit to 1, and perform a write operation, where the data written is the number plus 1, and the address written is pixel_data;

In response to detecting the end of the image frame, port B performs a read operation and reads the corresponding data in sequence starting from the initial address, and compares them, recording the maximum value data_max and address addr_max, where data_max is the number corresponding to the largest number of pixel values, and addr_max is the pixel value corresponding to the largest number of pixel values.

In the embodiment of the present application, it also includes:

In response to starting to receive pixel value data of an image, assigning the received pixel value pixel_data to the address of port B of the dual-port RAM, and at the same time setting the chip select of port B to 1, the read/write control position to 0, performing a read operation, and determining the number of pixel points whose pixel value is pixel_data according to the read result;

Set the chip select of port A of the dual-port RAM to 1, the read/write control bit to 1, and perform a write operation, where the data written is the number plus 1, and the address written is pixel_data;

In the embodiment of the present application, the adjustment coefficient is generated according to the number and pixel value corresponding to the largest number of pixel values, the total number of pixels of the image, and the pixel depth, further comprising:

In response to the fact that the number of pixel values with the largest number corresponds to a total number of pixels of the image that is less than a preset ratio, no peak clipping is performed and the adjustment coefficient is 0.

In the embodiment of the present application, it also includes:

In response to the total number of pixels of the image corresponding to the largest number of pixel values being not less than a preset ratio, peak clipping is performed, and an adjustment coefficient is determined according to a formula (data_max-a*Fream_S)*addr_max/2 ^pix_width , where a is a preset ratio.

In the embodiment of the present application, the value of a is 0.4.

In the embodiment of the present application, generating the adjustment coefficient according to the number and pixel value corresponding to the largest number of pixel values, the total number of pixels of the image, and the pixel depth includes:

When data_max is less than 0.4*Fream_S, no peak clipping calculation is performed and the adjustment coefficient Adjust is 0;

When data_max is not less than 0.4*Fream_S, the adjustment coefficient Adjust=(data_max-0.4*Fream_S)*addr_max/2 ^pix_width .

In the embodiment of the present application, the image is divided into a plurality of image blocks before the DCT transformation, and the pixel value of each pixel in each image block is added with the adjustment coefficient, further comprising:

The image is divided into a plurality of 8*8 image blocks, wherein each image block includes 64 pixels, and the pixel value of each pixel is added to the adjustment coefficient to obtain each adjusted image block.

In the embodiment of the present application, calculating the brightness expectation corresponding to each image block after adjustment and calculating the standard deviation corresponding to each image block based on the brightness expectation further includes:

The pixel values of the 64 pixels of the adjusted image block are added and averaged to obtain the expected brightness corresponding to the image block, and the difference between the pixel value of each of the 64 pixels and the expected brightness is calculated respectively, and then the 64 differences are accumulated to obtain the standard deviation corresponding to the image block.

In the embodiment of the present application, the pixel values of 64 pixels of the adjusted image block are added and averaged to obtain the brightness expectation corresponding to the image block, further comprising:

Shift the accumulated value to the right by the corresponding bit to obtain the expected brightness.

In the embodiment of the present application, comparing the standard deviation with the preset first standard deviation threshold and the second standard deviation threshold to determine the standard deviation coefficient of each image block further includes:

In response to the standard deviation being less than a first standard deviation threshold, the standard deviation coefficient is a maximum value among the 64 difference values;

In response to the standard deviation being not less than the first standard deviation threshold and less than the second standard deviation threshold, the standard deviation coefficient is a maximum value of the 64 difference values divided by the standard deviation;

In response to the standard deviation being not less than a second standard deviation threshold, the standard deviation coefficient is a minimum value among the 64 difference values.

In the embodiment of the present application, comparing the brightness expectation with the preset first brightness expectation threshold and the second brightness expectation threshold to determine the brightness expectation coefficient of each image block further includes:

In response to the brightness expectation being less than the first brightness expectation threshold, the brightness expectation coefficient is 1;

In response to the brightness expectation being not less than the first brightness expectation threshold and less than the second brightness expectation threshold, the brightness expectation coefficient is the brightness expectation divided by the first brightness expectation threshold;

In response to the brightness expectation being not less than the second brightness expectation threshold, the brightness expectation coefficient is the brightness expectation divided by the second brightness expectation threshold.

In the embodiment of the present application, updating the standard quantization parameter table using the expected brightness coefficient and the standard deviation coefficient corresponding to each image block further includes:

The standard quantization parameter table is multiplied by the conversion coefficient and then divided by the product of the brightness expectation coefficient and the standard deviation coefficient.

In the embodiment of the present application, the conversion coefficient is determined according to the image specification.

Based on the same concept, according to the second aspect, an embodiment of the present application further provides a system for improving JPEG compression quality, including:

A statistical module is configured to count pixel values in the image to be compressed to obtain the largest number of pixel values, and generate an adjustment coefficient according to the number and pixel value corresponding to the largest number of pixel values, the total number of pixels of the image, and the pixel depth;

a conversion module, configured to perform color mode conversion and DCT transformation on the image to obtain a coefficient matrix, wherein the image is divided into a plurality of image blocks before the DCT transformation, and the pixel value of each pixel point in each image block is adjusted by adding the adjustment coefficient;

A first calculation module is configured to calculate the brightness expectation corresponding to each adjusted image block and calculate the standard deviation corresponding to each image block based on the brightness expectation;

A comparison module is configured to compare the standard deviation with a preset first standard deviation threshold and a second standard deviation threshold to determine a standard deviation coefficient of each image block, and compare the brightness expectation with a preset first brightness expectation threshold and a second brightness expectation threshold to determine a brightness expectation coefficient of each image block;

The second calculation module is configured to obtain a standard quantization parameter table specified in the JPEG protocol and update the standard quantization parameter table using a brightness expectation coefficient and a standard deviation coefficient corresponding to each image block;

The quantization module is configured to quantize the coefficient matrix using the updated quantization parameter table to obtain a quantized coefficient matrix corresponding to each image block.

Based on the same concept, according to the third aspect, an embodiment of the present application further provides a chip, including a digital logic circuit, which implements the steps of the task scheduling method of any of the above embodiments when working.

Based on the same concept, according to a fourth aspect, an embodiment of the present application further provides an electronic device, comprising the above-mentioned chip.

The present application has one of the following beneficial technical effects: the method for improving JPEG compression quality proposed in the present application can optimize the encoding process image block by image block according to the characteristics of different video images, so that the part of the original image with rich information retains more detail information after encoding, while the part of the original image with less information and smoothly changing occupies fewer resources after encoding.

BRIEF DESCRIPTION OF THE DRAWINGS

In order to more clearly illustrate the embodiments of the present application or the technical solutions in the prior art, the drawings required for use in the embodiments or the description of the prior art will be briefly introduced below. Obviously, the drawings described below are only some embodiments of the present application. For ordinary technicians in this field, other embodiments can be obtained based on these drawings without paying any creative work.

FIG1 is a flowchart of an image compression process based on the JPEG protocol in the prior art;

FIG2 is a flow chart of a method for improving JPEG compression quality provided by an embodiment of the present application;

FIG3 is a flowchart of an image compression process based on the JPEG protocol provided by an embodiment of the present application;

FIG4 is a schematic diagram of the structure of a system for improving JPEG compression quality provided by an embodiment of the present application;

FIG5 is a schematic diagram of the structure of a computer device provided in an embodiment of the present application;

FIG6 is a schematic diagram of the structure of a computer non-volatile readable storage medium provided in an embodiment of the present application.

Detailed ways

In order to make the objectives, technical solutions and advantages of the present application more clearly understood, the embodiments of the present application are further described in detail below in combination with optional embodiments and with reference to the accompanying drawings.

It should be noted that all expressions using "first" and "second" in the embodiments of the present application are for distinguishing two non-identical entities with the same name or non-identical parameters. It can be seen that "first" and "second" are only for the convenience of expression and should not be understood as limitations on the embodiments of the present application. The subsequent embodiments will not explain this one by one.

According to one aspect of the present application, an embodiment of the present application proposes a method for improving JPEG compression quality, as shown in FIG2 , which may include the steps of:

S1, counting pixel values in the image to be compressed to obtain the largest number of pixel values, and generating an adjustment coefficient according to the number and pixel value corresponding to the largest number of pixel values, the total number of pixels of the image, and the pixel depth;

S2, performing color mode conversion and DCT transformation on the image to obtain a coefficient matrix, wherein the image is divided into a plurality of image blocks before the DCT transformation, and the pixel value of each pixel point in each image block is adjusted by adding the adjustment coefficient;

S3, calculating the brightness expectation corresponding to each image block after adjustment and calculating the standard deviation corresponding to each image block based on the brightness expectation;

S4, comparing the standard deviation with the preset first standard deviation threshold and the second standard deviation threshold to determine the standard deviation coefficient of each image block, and comparing the brightness expectation with the preset first brightness expectation threshold and the second brightness expectation threshold to determine the brightness expectation coefficient of each image block;

S5, obtaining a standard quantization parameter table specified in the JPEG protocol and updating the standard quantization parameter table using the brightness expectation coefficient and standard deviation coefficient corresponding to each image block;

S6, quantizing the coefficient matrix using the updated quantization parameter table to obtain a quantized coefficient matrix corresponding to each image block.

The method for improving JPEG compression quality proposed in this application can optimize the encoding process image block by image block according to the characteristics of different video images, so that the information-rich parts of the original image retain more detail information after encoding, while the parts of the original image with less information and smooth changes occupy fewer resources after encoding.

Call the preset dual-port RAM (Random Access Memory, where the data width of the dual-port RAM is log2Fream_S, the data depth is 2 ^pix_width , Fream_S is the total number of pixels of the image, and pix_width is the pixel depth;

In the embodiment of the present application, it also includes:

In response to starting to receive the pixel value data of the image, the received pixel value pixel_data is assigned to the address of the B port of the dual-port RAM, and the enable of the B port is set to 1, the read and write control position is set to 0, and a read operation is performed. The result determines the number of pixels whose pixel value is pixel_data;

Optionally, as shown in FIG3 , when counting pixel value histograms, hardware can be used for implementation. The format of image input is generally input in the order of each row of pixels from left to right and from top to bottom. First, a dual-port RAM is instantiated or called as the hardware carrier of the histogram information. The data width of the RAM is log2Fream_S, and the data depth is 2 ^pix_width , where Fream_S is the number of pixels of the image, such as the commonly used 1920*1080 resolution, the number of pixels is 2073600, then the data width of the RAM is 21 bits, and pix_width is the pixel depth, such as the commonly used 8bit (bit), then the data depth is 256.

The process of hardware implementing pixel value histogram statistics is as follows:

When the beginning of the image frame is detected, all data in the dual-port RAM are set to 0, which can be accomplished in hardware by using a reset or writing zeros to each address in turn.

When starting to receive valid data, the data value is recorded as pixel_data. First, pixel_data is assigned to the address ram_b_addr of the B port of the dual-port RAM, and the enable (or chip select) of the B port is set to 1, and the read-write control position is set to 0. The read operation is performed, and the result ram_b_dout, which represents the number of data values before the data is pixel_data, is added by 1 and assigned to the A port for a write operation, ram_a_din = ram_b_dout + 1, ram_a_addr = pixel_data, the enable (or chip select) of the A port is set to 1, and the read-write control position is set to 1. The pixel value of the received data is used as the address, and the number of the pixel value is used as the data written in the address.

When the end of the frame is detected, all image data information has been written into the dual-port RAM. At this time, the read operation is performed by port B, and the corresponding data can be read in sequence from the initial address to the maximum address, and compared in sequence, and the address and data of the largest value are recorded, recorded as data_max (the number corresponding to the largest number of pixel values) and addr_max (the pixel value corresponding to the largest number of pixel values).

In the embodiment of the present application, it also includes:

In the embodiment of the present application, the value of a is 0.4.

Optionally, when data_max＜0.4*Fream_S, no peak clipping calculation is performed, and the adjustment coefficient adjust is 0;

Next, when data_max≥0.4*Fream_S, Adjust=(data_max-0.4*Fream_S)*addr_max/2 ^pix_width .

Optionally, as shown in FIG3, after the original image is converted into a color mode, a Block (image block) is generated and stored, and the subsequent process is performed block by block. This part includes the 8x8 image block in FIG3, the write DDR (double rate synchronous dynamic random access memory) and the read DDR part. By calculating the address at the DDR, the image is stored in the DDR as 8x8 blocks, and is subsequently read out from the DDR as blocks. This part is not a key part of this proposal and will not be introduced in detail. It should be noted that the size of each image block is 8*8, and the number of image blocks that can be obtained from the original image is determined by the size of the original image.

Before performing the DCT transformation, the pixel value of each pixel in each image block is added with the adjustment coefficient, and then the DCT transformation is performed.

In this way, when DCT is calculated, the original data read out from DDR is added with the adjustment coefficient adjust as the input data of DCT calculation, that is, DCT_din=ddr_out+adjust.

In the embodiment of the present application, the pixel values of the 64 pixels of each image block after adjustment are added and averaged to obtain the brightness expectation corresponding to the image block, further comprising:

Optionally, the input data of DCT_din is transmitted to the quantization unit at the same time. First, the average value (that is, expectation) of the 8x8 block pixels is calculated. When implemented in hardware, 64 data are accumulated, and the accumulated result sum_block is shifted right by the corresponding bit, which is the brightness expectation value block_ave.

Then, the difference between the input data and the expected value is calculated in sequence, dif=DCT_din-block_ave, and the 64 differences are accumulated and recorded as the standard deviation dif_sum.

Optionally, the standard deviation can reflect the degree of dispersion of a set of data relative to the mean value. In images, it reflects the information richness contained in the image, and subjectively it is expressed as contrast. Images with large standard deviations often have high information richness, bright colors, obvious textures, and high contrast. On the contrary, the amount of information is insufficient. Therefore, the standard deviation can be used to calculate the factors of quantitative calculations.

The upper and lower truncation methods can be used to set a relatively large quantization coefficient for the part with too small standard deviation, because the partition itself does not have enough information. For the part with a lower level but greater than the lower limit, since it has enough information, it can be optimized. When the standard deviation is at a higher level, since the information itself is large enough, a smaller quantization coefficient is used to retain the detail information. Therefore, the standard deviation can be compared with the preset first standard deviation threshold and the second standard deviation threshold to determine the standard deviation coefficient of each image block:

When dif_sum＜dif_min, Y_dif_factor＝Y_dif_max;

When dif_min＜dif_sum＜dif_max, Y_dif_factor＝Y_dif_max/dif_sum;

When dif_sum＞dif_max, Y_dif_factor＝Y_dif_min.

That is, if the standard deviation is less than the first standard deviation threshold, the standard deviation coefficient is the maximum value among the 64 differences; if the standard deviation is not less than the first standard deviation threshold and less than the second standard deviation threshold, the standard deviation coefficient is the maximum value among the 64 differences divided by the standard deviation; if the standard deviation is not less than the second standard deviation threshold, the standard deviation coefficient is the minimum value among the 64 differences.

Optionally, similar to the calculation method of the standard deviation coefficient, two threshold parameters are first set. When the value is below the threshold Ymin, the standard brightness expectation coefficient 1 is used; when the value exceeds the threshold Ymax, as the brightness increases, the modulation coefficient becomes smaller, thereby suppressing the enhancement of the high brightness area while ensuring the normal enhancement of the lower brightness. Therefore, the brightness expectation and the preset first brightness expectation threshold and second brightness expectation threshold can be compared to determine the brightness expectation coefficient of each image block:

When block_ave＜Ymin, Y_ave_factor＝1;

When Ymin＜block_ave＜Ymax, Y_ave_factor＝block_ave/Ymin;

When block_ave＞Ymax, Y_ave_factor＝block_ave/Ymax;

That is, the brightness expectation is less than the first brightness expectation threshold, and the brightness expectation coefficient is 1; the brightness expectation is not less than the first brightness expectation threshold and less than the second brightness expectation threshold, and the brightness expectation coefficient is the brightness expectation divided by the first brightness expectation threshold; the brightness expectation is not less than the second brightness expectation threshold, and the brightness expectation coefficient is the brightness expectation divided by the second brightness expectation threshold.

Optionally, the jpeg protocol specifies a standard brightness and color quantization parameter table [Q_s] (essentially an 8*8 matrix). The standard quantization parameter table is then adjusted according to the calculated brightness expectation coefficient and standard deviation coefficient, that is, [Q] = [Q_s]*c/(Y_dif_factor*Y_ave_factor), that is, the standard quantization parameter table is multiplied by the conversion coefficient and divided by the product of the brightness expectation coefficient and the standard deviation coefficient.

Wherein, c is the conversion coefficient, which is a fixed value when the image specifications are determined.

Finally, the calculated quantization parameter table is updated to the quantization calculation process of image coding, and the corresponding division operation is performed to complete the quantization calculation.

The solution proposed in this application realizes global "peak clipping" processing and quantization coefficient calculation by calculating the global brightness histogram of the entire image, the standard deviation of each block, and the brightness expectation and other parameters, and then optimizes the quantization coefficient table block by block to achieve the best image encoding effect. Therefore, the method for improving JPEG compression quality proposed in this application can optimize the encoding process block by block according to the characteristics of different video images, so that the part of the original image with rich information retains more detailed information after encoding, and the part of the original image with less information and smooth changes occupies less resources after encoding.

Based on the same concept, according to another aspect of the present application, an embodiment of the present application further provides a system 400 for improving JPEG compression quality, as shown in FIG4 , comprising:

The statistics module 401 is configured to count the pixel values in the image to be compressed to obtain the largest number of pixel values, and The number and pixel value corresponding to the pixel value with the largest number, the total number of pixels of the image, and the pixel depth generate an adjustment coefficient;

The conversion module 402 is configured to perform color mode conversion and DCT transformation on the image to obtain a coefficient matrix, wherein the image is divided into a plurality of image blocks before the DCT transformation, and the pixel value of each pixel point in each image block is adjusted by adding the adjustment coefficient;

A first calculation module 403 is configured to calculate the brightness expectation corresponding to each image block after adjustment and calculate the standard deviation corresponding to each image block based on the brightness expectation;

The comparison module 404 is configured to compare the standard deviation with a preset first standard deviation threshold and a second standard deviation threshold to determine the standard deviation coefficient of each image block, and compare the brightness expectation with a preset first brightness expectation threshold and a second brightness expectation threshold to determine the brightness expectation coefficient of each image block;

The second calculation module 405 is configured to obtain a standard quantization parameter table specified in the JPEG protocol and update the standard quantization parameter table using the brightness expectation coefficient and the standard deviation coefficient corresponding to each image block;

The quantization module 406 is configured to quantize the coefficient matrix using the updated quantization parameter table to obtain a quantized coefficient matrix corresponding to each image block.

In the embodiment of the present application, the statistics module 401 is further configured to:

Call the preset dual-port RAM, where the data width of the dual-port RAM is log2Fream_S, the data depth is 2 ^pix_width , Fream_S is the total number of pixels of the image, and pix_width is the pixel depth;

Set the enable of port A of the dual-port RAM to 1, the read-write control position to 1, and perform a write operation, where the data written is the number plus 1, and the written address is pixel_data.

In the embodiment of the present application, the conversion module 402 is further configured to:

In the embodiment of the present application, the value of a is 0.4.

In the embodiment of the present application, the first calculation module 403 is further configured to:

The pixel values of the 64 pixels of the adjusted image block are added and averaged to obtain the brightness period corresponding to the image block. The standard deviation of the image block is obtained by accumulating the 64 differences and calculating the difference between the pixel value of each pixel and the expected brightness.

In the embodiment of the present application, the comparison module 404 is further configured to:

In the embodiment of the present application, the second calculation module 405 is further configured to:

The solution proposed in this application calculates the global brightness histogram of the entire image, the standard deviation of each block, the brightness expectation and other parameters, to achieve global "peak clipping" processing and quantization coefficient calculation, and then optimizes the quantization coefficient table block by block to achieve the best image encoding effect. Therefore, the method for improving JPEG compression quality proposed in this application can optimize the encoding process block by block according to the characteristics of different video images, so that the part of the original image with rich information retains more detailed information after encoding, and the part of the original image with less information and smooth changes occupies less resources after encoding.

Based on the same concept, according to another aspect of the present application, as shown in FIG5 , an embodiment of the present application further provides a chip 501, including:

The digital logic circuit 510 is included, and when the digital logic circuit 510 is working, the steps of the task scheduling method in any of the above embodiments are implemented.

Based on the same concept, according to another aspect of the present application, as shown in FIG. 6 , an embodiment of the present application further provides an electronic device 601 , comprising the above-mentioned chip 610 .

Finally, it should be noted that a person skilled in the art can understand that all or part of the processes in the method of the embodiment of the present application can be implemented by instructing the relevant hardware through a computer program, and the program can be stored in a computer non-volatile readable storage medium. When the program is executed, it may include the processes of the embodiments of the methods of the present application.

Furthermore, it should be appreciated that the computer non-volatile readable storage medium (eg, memory) herein may be either volatile memory or non-volatile memory, or may include both volatile memory and non-volatile memory.

Those skilled in the art will also appreciate that various exemplary logic blocks, modules, circuits and algorithm steps described in conjunction with the disclosure herein can be implemented as electronic hardware, computer software or a combination of the two. In order to clearly illustrate this interchangeability of hardware and software, a general description has been given to the functions of various schematic components, blocks, modules, circuits and steps. Whether this function is implemented as software or implemented as hardware depends on application and the design constraints imposed on the entire system. Those skilled in the art can implement the function in various ways for each application, but this implementation decision should not be interpreted as causing a departure from the disclosed scope of the present application embodiment.

The above are exemplary embodiments disclosed in the present application, but it should be noted that various changes and modifications may be made without departing from the scope disclosed in the embodiments of the present application as defined in the claims. The functions, steps and/or actions of the method claims according to the disclosed embodiments described herein do not need to be performed in any particular order. In addition, although the elements disclosed in the embodiments of the present application may be described or required in individual form, they may also be understood as multiple unless explicitly limited to the singular.

It should be understood that, as used herein, the singular forms "a", "an" are intended to include the plural forms as well, unless the context clearly supports an exception. It should also be understood that, as used herein, "and/or" refers to any and all possible combinations including one or more of the associated listed items.

The serial numbers of the embodiments disclosed in the above-mentioned embodiments of the present application are only for description and do not represent the advantages or disadvantages of the embodiments.

A person of ordinary skill in the art can understand that all or part of the steps of implementing the embodiments of the present application can be completed by hardware, or can be completed by instructing related hardware through a program, and the program can be stored in a computer non-volatile readable storage medium. The non-volatile readable storage medium mentioned above can be a read-only memory, a disk or an optical disk, etc.

A person of ordinary skill in the art should understand that the discussion of any embodiment of the present application is only exemplary and is not intended to imply that the scope of the disclosure of the embodiments of the present application (including the claims) is limited to these examples; under the idea of the embodiments of the present application, the technical features in the embodiments of the present application or in different embodiments can also be combined, and there are many other changes in different aspects of the embodiments of the present application as above, which are not provided in detail for the sake of simplicity. Therefore, any omissions, modifications, equivalent substitutions, improvements, etc. made within the spirit and principles of the embodiments of the present application should be included in the protection scope of the embodiments of the present application.

Claims

A method for improving the quality of Joint Photographic Experts Group JPEG compression, characterized by comprising the following steps:

Counting pixel values in the image to be compressed to obtain the largest number of pixel values, and generating an adjustment coefficient according to the number and pixel values corresponding to the largest number of pixel values, the total number of pixels of the image, and the pixel depth;

Performing color mode conversion and discrete cosine DCT transformation on the image to obtain a coefficient matrix, wherein the image is divided into a plurality of image blocks before the DCT transformation, and the pixel value of each pixel point in each image block is added with the adjustment coefficient to perform adjustment;

Calculating the adjusted brightness expectation corresponding to each image block and calculating the standard deviation corresponding to each image block based on the brightness expectation;

Comparing the standard deviation with a preset first standard deviation threshold and a second standard deviation threshold to determine a standard deviation coefficient of each image block, and comparing the brightness expectation with a preset first brightness expectation threshold and a second brightness expectation threshold to determine a brightness expectation coefficient of each image block;

Obtaining a standard quantization parameter table specified in the JPEG protocol and updating the standard quantization parameter table using the brightness expectation coefficient and the standard deviation coefficient corresponding to each image block;

The coefficient matrix is quantized using the updated quantization parameter table to obtain a quantized coefficient matrix corresponding to each image block.
The method according to claim 1, wherein counting the pixel values in the image to be compressed to obtain the largest number of pixel values further comprises:

Calling a preset dual-port random access memory RAM, wherein the data bit width of the dual-port RAM is log 2 Fream_S, the data depth is 2 pix_width , Fream_S is the total number of pixels of the image, and pix_width is the pixel depth;

In response to detecting the beginning of an image frame, all data in the dual-port RAM are set to 0.
The method according to claim 2, characterized in that setting all data in the dual-port RAM to 0 further comprises:

All data in the dual-port RAM are set to 0 by resetting.
The method according to claim 2, characterized in that setting all data in the dual-port RAM to 0 further comprises:

All data in the dual-port RAM are set to 0 by writing zero to each address in sequence.
The method according to claim 2, further comprising:

In response to starting to receive pixel value data of an image, assigning the received pixel value pixel_data to the address of the B port of the dual-port RAM, and setting the enable of the B port to 1, the read-write control position to 0, performing a read operation, and determining the number of pixel points whose pixel value is pixel_data according to the read result;

The enable of the A port of the dual-port RAM is set to 1, the read-write control bit is set to 1, and a write operation is performed, wherein the data written is the number plus 1, and the address written is pixel_data;

In response to detecting the end of the image frame, the B port performs a read operation and reads the corresponding data in sequence starting from the initial address, and compares them, recording the maximum value data_max and the address addr_max, where data_max is the number corresponding to the largest number of pixel values, and addr_max is the pixel value corresponding to the largest number of pixel values.
The method according to claim 2, further comprising:

In response to starting to receive pixel value data of an image, assigning the received pixel value pixel_data to the dual port

The address of port B of RAM, and at the same time set the chip select of port B to 1, the read and write control position to 0, perform a read operation, and determine the number of pixels whose pixel value is pixel_data according to the read result;

The chip select of the A port of the dual-port RAM is set to 1, the read-write control bit is set to 1, and a write operation is performed, wherein the data written is the number plus 1, and the address written is pixel_data;

In response to detecting the end of the image frame, the B port performs a read operation and reads the corresponding data in sequence starting from the initial address, and compares them, recording the maximum value data_max and the address addr_max, where data_max is the number corresponding to the largest number of pixel values, and addr_max is the pixel value corresponding to the largest number of pixel values.
The method of claim 5, wherein generating the adjustment coefficient according to the number and pixel value corresponding to the largest number of pixel values, the total number of pixels of the image, and the pixel depth further comprises:

In response to the fact that the number corresponding to the largest number of pixel values is smaller than a preset ratio of the total number of pixels of the image, no peak clipping is performed and the adjustment coefficient is 0.
The method according to claim 7, further comprising:

In response to the number of pixel values with the largest number corresponding to not less than a preset ratio of the total number of pixels of the image, peak clipping is performed, and an adjustment coefficient is determined according to a formula (data_max-a*Fream_S)*addr_max/2 pix_width , where a is a preset ratio.
The method according to claim 8, characterized in that the value of a is 0.4.
The method of claim 2, wherein generating the adjustment coefficient according to the number and pixel value corresponding to the largest number of pixel values, the total number of pixels of the image, and the pixel depth comprises:

When data_max is less than 0.4*Fream_S, no peak clipping calculation is performed and the adjustment coefficient Adjust is 0:

When data_max is not less than 0.4*Fream_S, the adjustment coefficient Adiust=(data_max-0.4*Fream_S)*addr_max/2 pix_width .
The method according to claim 1, characterized in that the image is divided into a plurality of image blocks before performing DCT transformation, and the pixel value of each pixel in each image block is added with the adjustment coefficient, further comprising:

The image is divided into a plurality of 8*8 image blocks, wherein each image block includes 64 pixels, and the pixel value of each pixel is added to the adjustment coefficient to obtain each adjusted image block.
The method according to claim 11, characterized in that calculating the adjusted brightness expectation corresponding to each image block and calculating the standard deviation corresponding to each image block based on the brightness expectation further comprises:

The pixel values of the 64 pixels of the adjusted image block are added and averaged to obtain the expected brightness corresponding to the image block, and the difference between the pixel value of each pixel in the 64 pixels and the expected brightness is calculated respectively, and then the 64 differences are accumulated to obtain the standard deviation corresponding to the image block.
The method of claim 12, wherein the pixel values of 64 pixels of the adjusted image block are added and then averaged to obtain the expected brightness corresponding to the image block, further comprising:

Shift the accumulated value to the right by the corresponding bit to obtain the expected brightness.
The method of claim 12, wherein comparing the standard deviation with a preset first standard deviation threshold and a second standard deviation threshold to determine a standard deviation coefficient of each image block further comprises:

In response to the standard deviation being less than a first standard deviation threshold, the standard deviation coefficient is a maximum value among the 64 difference values;

In response to the standard deviation being not less than a first standard deviation threshold and less than the second standard deviation threshold, the standard deviation coefficient is a maximum value of the 64 difference values divided by the standard deviation;

In response to the standard deviation being not less than the second standard deviation threshold, the standard deviation coefficient is a minimum value among the 64 difference values.
The method according to claim 1, characterized in that comparing the brightness expectation with a preset first brightness expectation threshold and a second brightness expectation threshold to determine a brightness expectation coefficient for each image block further comprises:

In response to the brightness expectation being less than a first brightness expectation threshold, the brightness expectation coefficient is 1;

In response to the brightness expectation being not less than a first brightness expectation threshold and less than the second brightness expectation threshold, the brightness expectation coefficient is the brightness expectation divided by the first brightness expectation threshold;

In response to the desired brightness being not less than the second desired brightness threshold, the desired brightness coefficient is the desired brightness divided by the second desired brightness threshold.
The method according to claim 1, characterized in that the standard quantization parameter table is updated using the brightness expectation coefficient and the standard deviation coefficient corresponding to each image block, further comprising:

The standard quantization parameter table is multiplied by a conversion coefficient and then divided by a product of the brightness expectation coefficient and the standard deviation coefficient.
The method according to claim 1, characterized in that the conversion coefficient is determined according to image specifications.
A system for improving JPEG compression quality, comprising:

a statistical module configured to count pixel values in the image to be compressed to obtain the largest number of pixel values, and generate an adjustment coefficient according to the number and pixel values corresponding to the largest number of pixel values, the total number of pixels of the image, and the pixel depth;

a conversion module, configured to perform color mode conversion and DCT transformation on the image to obtain a coefficient matrix, wherein the image is divided into a plurality of image blocks before the DCT transformation, and the pixel value of each pixel point in each image block is added with the adjustment coefficient for adjustment;

A first calculation module is configured to calculate the adjusted brightness expectation corresponding to each image block and calculate the standard deviation corresponding to each image block based on the brightness expectation;

A comparison module is configured to compare the standard deviation with a preset first standard deviation threshold and a second standard deviation threshold to determine a standard deviation coefficient of each image block, and compare the brightness expectation with a preset first brightness expectation threshold and a second brightness expectation threshold to determine a brightness expectation coefficient of each image block;

A second calculation module is configured to obtain a standard quantization parameter table specified in the JPEG protocol and update the standard quantization parameter table using the brightness expectation coefficient and the standard deviation coefficient corresponding to each image block;

The quantization module is configured to quantize the coefficient matrix using the updated quantization parameter table to obtain a quantized coefficient matrix corresponding to each image block.
A chip, characterized in that it includes a digital logic circuit, and when the digital logic circuit is in operation, the steps of the method described in any one of claims 1 to 17 are implemented.
An electronic device, characterized by comprising the chip as claimed in claim 19.