WO2022188070A1

WO2022188070A1 - Image encoding method and apparatus, and image decoding method and apparatus

Info

Publication number: WO2022188070A1
Application number: PCT/CN2021/080029
Authority: WO
Inventors: 牛兵兵; 赵文军; 邱孟品
Original assignee: 深圳市大疆创新科技有限公司
Priority date: 2021-03-10
Filing date: 2021-03-10
Publication date: 2022-09-15

Abstract

The present application discloses an image encoding method, comprising: acquiring first data to be encoded corresponding to an image to be encoded, wherein said first data comprises a plurality of pixel values of a plurality of pixels of a block in said image; from among the plurality of pixel values, determining a first minimum pixel value for the plurality of pixel values; according to the first minimum pixel value, determining a residual corresponding to each pixel value from among the plurality of pixel values; and compressing the residual corresponding to each pixel value, and the first minimum pixel value. By means of the image encoding and decoding methods disclosed in the present application, a relatively high precision can be retained for a dark part of an image, which conforms to the law that the sensing sensitivity of human eyes to a brightness change at a dark part is higher than that at a bright part, such that the compression performance is improved, and a good image quality is also ensured.

Description

Image coding method and device, and image decoding method and device

technical field

The present application relates to the technical field of image processing, and in particular, to an image encoding method, apparatus, and computer-readable storage medium, and also to an image decoding method, apparatus, and computer-readable storage medium.

Background technique

With the improvement of image resolution and the improvement of video frame rate, the scale of image data becomes larger and larger, and the transmission of image data will consume a lot of bandwidth resources. Especially for systems with relatively limited bandwidth resources, such as SOC (System-on-a-Chip) chip systems, insufficient bandwidth will directly affect the work efficiency of the system, so it is necessary to take measures to reduce the bandwidth resources required for image data transmission.

Encoding and compressing image data is an effective means to reduce bandwidth resources required for image data transmission, but the compression effect of existing image data encoding and decoding technologies still needs to be improved.

SUMMARY OF THE INVENTION

In view of this, embodiments of the present application provide an image encoding method, device, and computer-readable storage medium, and an image decoding method, device, and computer-readable storage medium, one of the purposes is to ensure that the image quality is within an acceptable range to maximize compression performance.

A first aspect of the embodiments of the present application provides an image encoding method, including:

acquiring first data to be encoded corresponding to the image to be encoded, the first data to be encoded includes multiple pixel values of multiple pixels of a block in the image to be encoded;

determining a first minimum pixel value for the plurality of pixel values from the plurality of pixel values;

determining, according to the first minimum pixel value, a residual corresponding to each pixel value in the plurality of pixel values;

The residual corresponding to each pixel value and the first minimum pixel value are compressed.

A second aspect of the embodiments of the present application provides an image decoding method, the method includes:

obtaining the first to-be-decoded data corresponding to the to-be-decoded image;

decompressing the first data to be decoded to determine a first minimum pixel value and a plurality of residuals;

From the first minimum pixel value and the plurality of residuals, a plurality of pixel values for a plurality of pixels of the decoded image are determined.

A third aspect of the embodiments of the present application provides an image encoding apparatus, including: a processor and a memory storing a computer program, where the processor implements the following steps when executing the computer program:

A fourth aspect of an embodiment of the present application provides an image decoding apparatus, including: a processor and a memory storing a computer program, where the processor implements the following steps when executing the computer program:

A fifth aspect of the embodiments of the present application provides a computer-readable storage medium, where the computer-readable storage medium stores a computer program, and when the computer program is executed by a processor, implements the image encoding method provided by the embodiments of the present application.

A sixth aspect of the embodiments of the present application provides a computer-readable storage medium, where the computer-readable storage medium stores a computer program, and when the computer program is executed by a processor, implements the image decoding method provided by the embodiments of the present application.

In the image coding method provided by the embodiment of the present application, the coding end may determine the residual corresponding to each pixel value according to the first minimum pixel value corresponding to each pixel value, and may compress the residual corresponding to each pixel value and the first minimum pixel value Then, it is transmitted to the decoding end, so that the decoding end can reconstruct and obtain each pixel value based on the residual corresponding to each pixel value and the first minimum pixel value. It can be seen that, because only the first minimum pixel value and the residual corresponding to each pixel value need to be transmitted between the encoding end and the decoding end, and the original pixel value of each pixel does not need to be transmitted, the amount of data to be transmitted is reduced. In addition, since the residual corresponding to each pixel value is determined based on the first minimum pixel value, the pixel value (dark part of the image) close to the first minimum pixel value can retain high precision, which is consistent with the human eye's sensitivity to changes in brightness. The perceptual sensitivity is consistent with the law that the dark part is higher than the bright part, which improves the compression performance while ensuring good image quality.

Description of drawings

In order to illustrate the technical solutions in the embodiments of the present application more clearly, the following briefly introduces the drawings that are used in the description of the embodiments. Obviously, the drawings in the following description are only some embodiments of the present application. For those of ordinary skill in the art, other drawings can also be obtained from these drawings without creative labor.

FIG. 1 is a flowchart of an image coding method provided by an embodiment of the present application.

FIG. 2 is a graph of logarithmic transformation provided by an embodiment of the present application.

FIG. 3 is a layout format diagram of an original raw image provided by an embodiment of the present application.

FIG. 4 is a graph of logarithmic transformation under specific parameters provided by an embodiment of the present application.

FIG. 5 is a schematic diagram of block division provided by an embodiment of the present application.

FIG. 6 is a flowchart of an image decoding method provided by an embodiment of the present application.

FIG. 7 is a schematic structural diagram of an image encoding apparatus provided by an embodiment of the present application.

FIG. 8 is a schematic structural diagram of an image decoding apparatus provided by an embodiment of the present application.

Detailed ways

The technical solutions in the embodiments of the present application will be clearly and completely described below with reference to the drawings in the embodiments of the present application. Obviously, the described embodiments are only a part of the embodiments of the present application, but not all of the embodiments. Based on the embodiments in the present application, all other embodiments obtained by those of ordinary skill in the art without creative efforts shall fall within the protection scope of the present application.

Encoding and compressing image data is an effective means to reduce the bandwidth resources required for image data transmission, but some image data encoding and decoding technologies do not consider the human eye's perception sensitivity to brightness changes, and the compression effect is not ideal.

To this end, an embodiment of the present application provides an image encoding method, and reference may be made to FIG. 1 , which is a flowchart of the image encoding method provided by the embodiment of the present application. The method can be applied to the encoding end, which includes the following steps:

S102: Acquire first data to be encoded corresponding to the image to be encoded.

S104. Determine a first minimum pixel value from a plurality of pixel values of the first data to be encoded.

S106. Determine, according to the first minimum pixel value, a residual corresponding to each pixel value in the plurality of pixel values.

S108: Compress the residual corresponding to each pixel value and the first minimum pixel value.

The first data to be encoded may include multiple pixel values of multiple pixels in a block of the image to be encoded. Since the pixel value can represent the brightness, the pixel position with a small pixel value can correspond to the dark part in the image), and the pixel value is large. The pixel positions of can correspond to bright parts in the image.

The human eye's perceptual sensitivity to changes in brightness is different in dark and bright parts. Specifically, the human eye has a strong perceptual sensitivity to the brightness change of the dark part, that is, even a small brightness change in the dark part is easy to be detected by the human eye, while the human eye's perceptual sensitivity to the brightness change of the bright part is Relatively low, that is, it is not sensitive to the brightness change of the bright part, and has a high tolerance to the brightness error of the bright part. Therefore, in the process of image compression, high precision can be retained for the dark parts of the image, and the compression strength can be appropriately increased for the bright parts of the image, so that a balance between image quality and compression performance can be achieved. Guaranteed good picture quality.

In one embodiment, the compression rate of the residual corresponding to the pixel value may be positively correlated with the size of the residual corresponding to the pixel value, that is, the larger the residual, the higher the compression rate of the residual, and the smaller the residual. The lower the compression ratio of this residual. Since the pixel value of the pixel in the dark part of the image corresponds to a small residual, and the small residual corresponds to a low compression rate, the residual corresponding to the pixel in the dark part of the image can be encoded with a low compression rate, so that the pixel in the dark part can be encoded with a low compression rate. Residuals corresponding to points can retain high accuracy. For the pixels in the bright part of the image, because the pixel value of the pixel in the bright part has a large residual error, and the large residual error corresponds to a large compression rate, a higher compression rate can be used for the pixel point in the bright part. The corresponding residual is encoded, which reduces the amount of data transmitted on the basis of ensuring image quality.

In one embodiment, the compression of the residual corresponding to each pixel value may be implemented by logarithmic transformation. The logarithmic transformation can realize the mapping of values, that is, one value can be mapped to another value. Referring to FIG. 2 , FIG. 2 is a graph of logarithmic transformation provided by an embodiment of the present application. It can be seen that the mapping relationship of the logarithmic transformation is similar to the logarithmic curve in the shape of the curve, that is, the slope of the curve tends to be flat as the independent variable increases. Therefore, the independent variable interval in the first half of the curve can be mapped to a larger The range of the independent variable in the second half can be mapped to a smaller range. It can be understood in conjunction with Figure 2. As shown in Figure 2, the first half of the independent variable interval [a, b] can be mapped to the interval [A, B], and the second half of the independent variable interval of the same length as [a, b] [c,d] can be mapped to the interval [C,D] which is significantly smaller in length than [A,B].

Based on the above characteristics of logarithmic transformation, the residuals corresponding to each pixel value are compressed by logarithmic transformation, so that the residuals with smaller values can retain higher precision after mapping, and the residuals with larger values can be larger. The compression of strength conforms to the law that the human eye's perception sensitivity to changes in brightness is higher in the dark part than in the bright part. In addition, since the larger the light-sensing value of the camera Sensor (ie the sensor), the larger the noise will be, so mapping the residual with a larger value to a smaller range can also remove the noise to a certain extent.

In one embodiment, compressing the data may be to map the value of the data from the original range to a target range with a smaller overall value, and to realize this mapping, in one example, the value may be first converted into Its proportion in the original range, and then complete the final mapping according to the proportion and the target range. For the compression of the residual corresponding to each pixel value, here, the first residual can be used as the compression object for description, the first residual can be the residual corresponding to the first pixel value, and the first pixel value can be the first to-be-coded Any pixel value in the data.

When the first residual is compressed by logarithmic transformation, in one embodiment, the minimum residual range to which the first residual belongs may be determined, and the first target number of bits occupied by the first residual after compression may be determined , and compress the first residual according to the minimum residual range and the first target number of bits. Here, the minimum residual range to which the first residual belongs can be understood in combination with the original range described above, which can be used to convert the first residual to its proportion in the minimum residual range. The first target number of bits can be understood in combination with the target range described above. For example, when the first target number of bits is 5, it means that the first residual is compressed to 5 bits, then the compressed first residual The maximum value cannot exceed 31 (2 ⁵ -1), that is, the corresponding target range is [0,31].

It should be noted that the reason why the minimum residual error range to which the first residual belongs is determined, rather than any residual error range to which the first residual belongs, is to enable the first residual to retain higher accuracy after compression. Since the residual range to which the first residual belongs is used to convert the first residual to its corresponding proportion in the residual range, the residual range needs to match the value of the first residual, otherwise the compressed value will be reduced. The precision of the first residual. For example, if the value of the first residual is 18, the residual range to which the first residual belongs can be [0,31] (corresponding to 5 bits) or [0,63] (corresponding to 6 bits) , but the minimum residual range [0, 31] matches the first residual better, which can better preserve the accuracy of the compressed first residual.

When determining the minimum residual error range to which the first residual belongs, in one embodiment, the minimum residual error range corresponding to the block may be determined as the minimum residual error range to which the first residual belongs. Here, the first data to be encoded may correspond to a block in the image to be encoded, the first pixel value corresponding to the first residual may belong to the block, and it may be determined that the residuals corresponding to the pixel values of the block together fall into (ie, the minimum residual range corresponding to the block), and the minimum residual range corresponding to the block may be used as the minimum residual range to which the first residual belongs. When determining the minimum residual error range corresponding to the block, in one embodiment, the maximum residual error may be determined from the residual errors corresponding to each pixel value of the block, and the minimum residual error range is determined according to the maximum residual error. For example, if a block can be a 16*1 data block, including 16 pixel values, the maximum residual difference diffmax among the residuals corresponding to the 16 pixel values can be determined, and the maximum residual difference diffmax can be determined according to the maximum residual difference. Determine the minimum residual range diffLD corresponding to the block.

Considering that when the block is too large, there may be a large gap between the residuals corresponding to different pixel values in the block, resulting in a low degree of adaptation between the minimum residual range corresponding to the block and the first residual. The value of a residual may be 7, and the most suitable residual range of the first residual is [0,7] (corresponding to 3 bits), but the maximum residual of the residual corresponding to each pixel value in the block may be is 30, then the minimum residual range determined according to the maximum residual may be [0, 31]. Obviously, the minimum residual range corresponding to this block does not have a high degree of adaptation to the first residual. The residual range compresses the first residual, which will result in a larger reconstruction error of the first residual and affect the image quality.

Therefore, in one embodiment, the block can be divided into a plurality of sub-blocks, and the first pixel value corresponding to the first residual can belong to the first sub-block of the plurality of sub-blocks, then the The minimum residual range corresponding to a sub-block is used as the minimum residual range to which the first residual belongs. Since the sub-block contains fewer pixel values than the block, the difference between the residuals corresponding to different pixel values in the sub-block has a high probability to be reduced, so that the minimum residual value corresponding to the first sub-block will be reduced. The difference range has a higher probability and can better fit the first residual. When determining the minimum residual error range corresponding to the first subblock, the maximum residual error may be determined according to the residual error corresponding to each pixel value in the first subblock, and the minimum residual error range may be determined according to the maximum residual error. For example, a 16*1 block can be divided into two 8*1 sub-blocks, then for the first sub-block, the residuals corresponding to the 8 pixel values in the first sub-block can be determined. The maximum residual difference diffmax in , and the minimum residual error range diffLD corresponding to the first sub-block can be determined according to the maximum residual difference diffmax.

Further, since the residual corresponding to each pixel value is determined based on the first minimum pixel value of each pixel value in the block, after the block is divided into a plurality of sub-blocks, the first minimum pixel value only exists therein. In one sub-block, that is, for other sub-blocks, the actual minimum pixel value is not necessarily equal to the first minimum pixel value. Therefore, in one embodiment, a second minimum pixel value of the plurality of pixel values in the first subblock may be determined, and a minimum value difference between the second minimum pixel value and the first minimum pixel value may be determined , the minimum residual error range corresponding to the first sub-block can be adjusted according to the minimum difference value, and the adjusted minimum residual error range can be re-determined as the minimum residual error range corresponding to the first sub-block. By adjusting the minimum residual error range corresponding to the first sub-block, the compression accuracy of the first residual error can be further improved.

When adjusting the minimum residual error range corresponding to the first sub-block according to the minimum value difference, in one embodiment, the adjustment amount corresponding to the minimum value difference may be determined according to the first correspondence, so that The minimum residual error range corresponding to the first sub-block is adjusted according to the adjustment amount. The first correspondence may be the correspondence between the minimum difference value, the adjustment amount, and the first index value. In an example, the first correspondence may be the first table, and the first table may be recorded as adjTable. Among them, the minimum difference value may correspond to the adjustment amount, and the adjustment amount may correspond to the first index value adindex.

In the above, when the minimum residual error range is determined according to the maximum residual error, it may be determined according to the second corresponding relationship. Specifically, the second correspondence may be the correspondence between the maximum residual error, the minimum residual error range, and the second index value. For example, the second correspondence may be a second table, and the second table may be recorded as diffLDTable, in diffLDTable , the maximum residual difference diffmax may correspond to the minimum residual error range diffLD, and the minimum residual error range diffLD may correspond to the second index value diffindex.

In one embodiment, after the adjustment amount corresponding to the minimum difference value is determined according to the first correspondence, the first index value corresponding to the adjustment amount may also be determined, and the first index value may be written into the code stream and transmitted to the decoding end. After the minimum residual error range is determined according to the maximum residual error, the second index value corresponding to the minimum residual error range may also be determined, and the second index value may be written into the code stream and transmitted to the decoding end. Then, the decoding end can decode the code stream to obtain the first index value adindex and the second index value diffindex, and can determine the minimum residual error range corresponding to the first sub-interval by the following formula:

diffLD=diffLDTable[diffindex]-adjTable[adindex].

The above diffLDTable table and adjTable table can be obtained by pre-configuration. For the diffLDTable table, in an implementation manner, the maximum residual error range may be arranged in increments of the second index value, that is, the maximum residual error range corresponding to the last second index value may be set to the maximum value of the corresponding number of bits, but You can also make appropriate adjustments according to the real scene.

The first target number of bits occupied by the first residual after compression may be determined by allocation of the number of bits. In an implementation manner, the encoding of the image data in this embodiment of the present application may adopt a fixed-length encoding manner, that is, the number of bits occupied by the image data of a specified scale after encoding may be fixed. In this way, on the basis of determining the total number of bits, the smallest block to which the first pixel value belongs (if the block is not divided, the smallest block is the block itself, if the block is divided, then The smallest block is the largest residual corresponding to the first sub-block where the first pixel value is located, and initial allocation of the number of bits is performed on the first residual. In one example, the maximum residual corresponding to the minimum block is positively correlated with the number of bits available for the first residual in the initial allocation, that is, the larger the maximum residual corresponding to the minimum block, the greater the number of bits in the minimum block. The residuals corresponding to the pixel values of , can be allocated more bits. In one embodiment, since the largest residual can correspond to the second index value based on the second correspondence, the first residual can also be mapped to the first residual according to the second index value corresponding to the smallest block to which the first pixel value belongs. The initial allocation of the number of bits is performed.

Further, if there are still bits remaining after the initial allocation, the remaining bits can be preferentially allocated to the residuals corresponding to the pixel values at the division boundary. Here, the pixel value at the division boundary may be a pixel value adjacent or close to the division or division operation, for example, it may be the pixel value at the boundary of each sub-block, or the pixel at the boundary of the block. value. By preferentially allocating the remaining bits to the residuals corresponding to the pixel values at the division boundary, the pixel values at the division boundary can have higher reconstruction accuracy, thereby reducing the blocking effect caused by block division.

After the initial allocation of the number of bits and the allocation of the remaining number of bits, the number of bits obtained from the initial allocation and the number of bits obtained from the remaining allocation are added to determine the first target bit corresponding to the first residual. digits. When compressing the first residual according to the minimum residual range and the first target number of bits, specifically, the first residual can be converted into a proportion of the minimum residual range, and the proportion can be passed through gamma Transform and compress to a value range corresponding to the first target number of bits. You can refer to the following first formula, which is an expression for compressing the residual by logarithmic transformation:

Wherein, diff can represent the first residual to be compressed, diffLD can represent the minimum residual range to which the first residual belongs,

It can represent the proportion corresponding to the first residual in the minimum residual range, and the diffdepth can represent the first target number of bits corresponding to the first residual. It can be seen that the above formula as a whole corresponds to the basic form of gamma transformation, and the basic form of gamma transformation can be expressed as:

out=C*power(a,g)

Among them, power(a, g) represents the g power of a, and g can represent the gamma parameter value. 2 ^diffdepth -1 may represent a numerical range corresponding to the first target number of bits.

Referring to FIG. 2, it can be seen that if the value to be compressed by logarithmic transformation is very small, the value may be larger after compression by logarithmic transformation than before compression, which has the opposite effect. Therefore, in one embodiment, If the residual after compression is larger than the residual before compression, the value of the residual after compression may be modified to the value of the residual before compression, which corresponds to the second row of the above-mentioned first formula.

Considering that the decoding end reconstructs each pixel value based on the reconstructed first minimum pixel value and the residual, therefore, in order to reduce the reconstruction error, the coding end can compress the first minimum pixel value after compressing the compressed first pixel value. The minimum pixel value is reconstructed (for example, it can be reconstructed by inverse logarithmic transformation) to obtain the reconstructed first minimum pixel value, so that when calculating the residual corresponding to each pixel value, each pixel value and the reconstructed first minimum value can be calculated. The residual between the pixel values (which can be regarded as the corrected residual) can be compressed and transmitted to the decoding end after the residual determined based on the reconstructed first minimum pixel value.

When compressing the first minimum pixel value, in one embodiment, a residual between each pixel value and the first minimum pixel value may be determined (the residual here is the difference between each pixel value and the original first minimum pixel value) The actual residual between pixel values), and the maximum residual among the residuals corresponding to each pixel value can be determined, so that the theoretical maximum value of the first minimum pixel value can be determined according to the maximum residual, according to the theoretical maximum value The second target number of bits occupied by the first minimum pixel value after compression may be determined, and the first minimum pixel value may be compressed according to the second target number of bits. For ease of understanding, for example, the number of bits corresponding to the pixel value is 12, that is, the theoretical maximum value of the pixel value can be 4095 (2 ¹² -1). If the residual between each pixel value and the first minimum pixel value is The maximum residual error in is 2500, then it can be determined that the theoretical maximum value of the first minimum pixel value is 1595, that is, 11 bits are enough to represent the first minimum pixel value, so it can be determined that the first minimum pixel value corresponds to the first minimum pixel value. The target number of bits is 11, which can compress the first minimum pixel value from the original 12 bits to 11 bits.

In one embodiment, the compression of the first minimum pixel value may also be performed by a logarithmic transformation. The compressed first minimum pixel value may be transmitted to the decoding end, and the pixel position corresponding to the first minimum pixel value may also be written into the code stream and transmitted to the decoding end.

In one embodiment, the image to be encoded may be an original raw image. The original raw image may be an image in a bayer format, as shown in FIG. 3 , which is a layout format diagram of the original raw image provided by the embodiment of the present application. Figure 3 shows four arrangements of (a), (b), (c), and (d) of the original raw image, which include pixels corresponding to different color components, such as R, Gr, Gb, B four pixels of each color component. Considering that there is a strong correlation between the pixels of the same color component, the redundancy is also large, therefore, in an implementation manner, the original encoded data block of the to-be-encoded image may be de-interleaved to obtain the first color component. Corresponding encoded data block. For example, the original encoded data block of the to-be-encoded image can be a 128*1 data block, that is, a row of pixel values corresponding to one row (here is only an example, when memory resources are sufficient, an original encoded data block can also include multiple rows of pixel values. ), after de-interleaving the original encoded data block, two encoded data blocks corresponding to different color components of 64*1 can be obtained. Referring to FIG. 3 , when the original encoded data block corresponds to a row of pixels, the original encoded data block only includes pixel values corresponding to two color components. Here, the specific operation of de-interleaving may be to organize the pixels in the original raw image according to the color components to which they belong.

After de-interleaving to obtain the encoded data block corresponding to the first color component (here, the first color component may be any color component), logarithmic transformation may be used for each pixel value in the encoded data block corresponding to the first color component. compression. The compression here can reduce the size of the pixel value as a whole and improve the compression performance. Among them, the expression of logarithmic transformation has the same form as the first formula mentioned above, and can be expressed as follows:

Among them, in represents the original pixel value of the input, and inmax is the maximum value corresponding to the number of bits occupied by the original pixel value. For example, the number of bits occupied by the original pixel value is 14, then inmax can be 16383 (2 ^14-1 ), bl (Black level) is the offset value, so that the original pixel value can be further saved by reducing the offset value and then converting it into a proportion, and A is the bit occupied after compression The maximum value corresponding to the number of bits can be, for example, 4095 (2 ¹² -1), that is, the compression target is to compress the original pixel value of each pixel to 12 bits.

The logarithmic transformation can also be understood in conjunction with the relevant descriptions in the foregoing, and will not be repeated here. Reference may also be made to FIG. 4 , which is a graph of logarithmic transformation under specific parameters provided by an embodiment of the present application, where the specific parameters include: bl=512, inmax=16383, A=4095, and g=0.45.

After compressing the pixel values in the encoded data block corresponding to the first color component, the first to-be-encoded data block can be obtained. For ease of understanding, the above example can be continued. The encoded data block corresponding to the first color component is a 64*1 data block, and the first to-be-encoded data block obtained by compressing it is still a 64*1 data block. In an implementation manner, the first data block to be encoded may be divided to obtain a plurality of first data to be encoded. Referring to FIG. 5, FIG. 5 is a schematic diagram of block division provided by an embodiment of the present application. If the first data block to be encoded is called a block and the first data to be encoded is called a slice, in the example shown in FIG. A block (64*1 data block) can be divided into 4 slices, that is, each slice can correspond to a 16*1 data block. After a plurality of first data to be encoded are obtained by division, each of the first data to be encoded may be processed separately, and the specific processing method can be described above, and will not be repeated here.

In one embodiment, by performing multiple division operations on the original encoded data block, including the operation of dividing the block into slices and the operation of dividing the block into sub-blocks mentioned above, it is possible to avoid local unevenness causing the overall block. effect, thereby improving the quality of the reconstructed image.

As mentioned above, when the allocation of the number of bits (ie, the allocation of the code rate) is performed, the remaining number of bits may be preferentially allocated to the residual corresponding to the pixel value at the division boundary. Here, the pixel values at the division boundaries may further include pixel values at the boundaries of the respective first data blocks to be encoded and pixel values at the boundaries of the first data blocks to be encoded.

In addition, in an implementation manner, the method provided by the embodiment of the present application takes into account the characteristics of the original raw image in the bayer format, de-interleaves the original raw image, and encodes and decodes the encoded data blocks corresponding to each color component. The original raw images in the format also have better compression performance.

Referring to FIG. 6 below, FIG. 6 is a flowchart of an image decoding method provided by an embodiment of the present application. The method can be applied to the decoding end, and the method can include:

S602: Obtain first data to be decoded corresponding to the image to be decoded.

S604. Decompress the first data to be decoded to determine a first minimum pixel value and multiple residuals.

S606. Determine multiple pixel values of multiple pixels of the decoded image according to the first minimum pixel value and the multiple residuals.

The decoding end can decompress the first data to be decoded to obtain the first minimum pixel value and a plurality of residuals, where the residual obtained by decoding is determined based on the first minimum pixel value, so according to the plurality of residuals and The first minimum pixel value can be reconstructed to obtain multiple pixel values of the target image.

In one embodiment, the compression ratio of the residual corresponding to the pixel value is positively correlated with the size of the residual corresponding to the pixel value.

In an embodiment, the data to be decoded includes the plurality of compressed residuals, and the decompressing the data to be decoded includes:

The compressed residuals are decompressed by inverse log transformation.

In an embodiment, the decompression of the compressed residuals is performed by inverse logarithmic transformation, including:

The compressed first residual is decompressed according to the minimum residual range to which the first residual belongs and the first target number of bits occupied by the first residual after compression, wherein the first residual is decompressed. A residual is the residual corresponding to the first pixel value, and the first pixel value is any pixel value in the block.

In an embodiment, according to the minimum residual range to which the first residual belongs and the first target number of bits occupied by the first residual after compression, the compressed first residual is Unzip, including:

According to the first target number of bits, the compressed first residual is decompressed by inverse gamma transform to obtain the proportion of the first residual corresponding to the minimum residual range;

The first residual is determined according to the proportion and the minimum residual range.

It can be understood that since the encoding end compresses the residual through logarithmic transformation, the decoding end can decompress the compressed residual through inverse logarithmic transformation. The decompression process is the opposite of the compression process, and the calculation formula corresponding to the decompression can be derived in combination with the first formula provided above.

In an embodiment, the minimum residual error range to which the first residual belongs is determined according to a second correspondence relationship and a second index value, and the second correspondence relationship is a pre-configured minimum residual error range and the second The corresponding relationship between index values, where the second index value is obtained by decoding the data to be decoded.

It can be understood that if the coding segment does not divide the block into sub-blocks, the decoding end can use the second index value decoded in the code stream and the second corresponding relationship preconfigured by the local end that is the same as that of the coding end, to Determines the minimum residual range to which the first residual belongs.

In one embodiment, the minimum residual range to which the first residual belongs is determined in the following manner:

determining the minimum residual error range according to the second correspondence and the second index value;

Determine the adjustment amount corresponding to the minimum residual error range according to the first correspondence and the first index value;

Adjust the minimum residual error range according to the adjustment amount, and determine the adjusted minimum residual error range as the minimum residual error range to which the first residual belongs;

Wherein, the first correspondence is the correspondence between the pre-configured adjustment amount and the first index value, the second correspondence is the correspondence between the pre-configured minimum residual error range and the second index value, the first Both the index value and the second index value are decoded from the data to be decoded.

If the coding segment divides the block into sub-blocks, the decoding end can use the first index value and the second index value decoded in the code stream, according to the first index value and the first corresponding relationship pre-configured by the local end. , the adjustment amount can be determined, and according to the second index value and the second correspondence, the minimum residual error range to which the first residual belongs can be determined, so that the minimum residual error range can be adjusted according to the adjustment amount to determine the first residual error The minimum residual range to which it ultimately belongs.

In one embodiment, the determining, according to the first minimum pixel value and the plurality of residuals, a plurality of pixel values for a plurality of pixels of the image to be decoded includes:

Decompressing is performed according to the first minimum pixel value and the plurality of residuals to obtain a plurality of pixel values of a plurality of pixels of the image to be decoded.

In an embodiment, the performing decompression according to the first minimum pixel value and the plurality of residuals includes:

Decompression is performed by inverse log transform based on the first minimum pixel value and the plurality of residuals.

In the image decoding method provided in the embodiment of the present application, the decoding end can decompress the first data to be decoded to obtain the first minimum pixel value and multiple residuals, and can, according to the multiple residuals and the first minimum pixel value, Multiple pixel values of the resulting image can be reconstructed. It can be seen that, because only the first minimum pixel value and the residual corresponding to each pixel value need to be transmitted between the encoding end and the decoding end, it is not necessary to transmit the original pixel value of each pixel, thus reducing the amount of data transmitted. Moreover, each pixel value is reconstructed based on the first minimum pixel value, so the pixel value (dark part of the image) close to the first minimum pixel value can retain high precision, which is related to the human eye's perception sensitivity to brightness changes. The dark part is higher than the bright part in accordance with the law, which can improve the compression performance and ensure good image quality.

It should be noted that, in an implementation manner, the image encoding method and the image decoding method provided by the embodiments of the present application may be applied to the process in which the camera sensor transmits the collected raw raw image to an image processing chip such as an ISP.

Referring to FIG. 7 below, FIG. 7 is a schematic structural diagram of an image encoding apparatus provided by an embodiment of the present application. The apparatus includes: a processor 710 and a memory 720 storing a computer program, the processor implements the following steps when executing the computer program:

Optionally, the compression ratio of the residual corresponding to the pixel value is positively correlated with the size of the residual corresponding to the pixel value.

Optionally, when compressing the residual corresponding to each pixel value, the processor is configured to compress the residual corresponding to each pixel value by logarithmic transformation.

Optionally, when compressing the residual corresponding to each pixel value by logarithmic transformation, the processor is used to determine the minimum residual range to which the first residual belongs, where the first residual is The residual corresponding to the first pixel value, the first pixel value is any pixel value in the block; determine the first target number of bits occupied by the first residual after compression; according to the minimum The residual range and the first target number of bits compress the first residual.

Optionally, the processor is configured to convert the first residual into the minimum residual when compressing the first residual according to the minimum residual range and the first target number of bits. The ratio of the residual range; the ratio is compressed into the numerical range corresponding to the first target bit number through gamma transformation.

Optionally, the minimum residual range to which the first residual belongs is the minimum residual range corresponding to the block, and the processor is used to determine the minimum residual range corresponding to the block when determining the minimum residual range corresponding to the block. The maximum residual of the residual corresponding to each pixel value in the block; the minimum residual range is determined according to the maximum residual.

Optionally, the block is divided into multiple sub-blocks, the first pixel value belongs to the first sub-block of the multiple sub-blocks, and the minimum residual range to which the first residual belongs is: the minimum residual error range corresponding to the first sub-block, when the processor determines the minimum residual error range corresponding to the first sub-block, to determine each pixel value in the first sub-block The corresponding maximum residual error of the residual error; the minimum residual error range is determined according to the maximum residual error.

Optionally, the processor is further configured to determine a second minimum pixel value of a plurality of pixel values in the first subblock; according to the difference between the second minimum pixel value and the first minimum pixel value; The minimum residual error range corresponding to the first sub-block is adjusted to determine the adjusted minimum residual error range for the first sub-block.

Optionally, the processor adjusts the minimum residual error range corresponding to the first sub-block according to the minimum value difference between the second minimum pixel value and the first minimum pixel value, When determining the adjusted residual error range for the first sub-block, the adjustment amount corresponding to the minimum difference value is determined according to the first correspondence, and the first sub-block is adjusted according to the adjustment amount. The minimum residual error range corresponding to the block is adjusted, and the first correspondence is the correspondence between the pre-configured minimum difference value and the adjustment amount.

Optionally, the first corresponding relationship is the corresponding relationship between the minimum difference value, the adjustment amount and the first index value, and the processor is further configured to determine, according to the first corresponding relationship, the first corresponding relationship of the adjustment amount. An index value, the first index value is used for transmission to the decoding end.

Optionally, when determining the minimum residual error range according to the maximum residual error, the processor is configured to determine, according to a second correspondence relationship, a minimum residual error range corresponding to the maximum residual error, and the second correspondence relationship is the correspondence between the preconfigured maximum residual and the minimum residual range.

Optionally, the second correspondence is a correspondence between the maximum residual, the minimum residual range and the second index value, and the processor is further configured to determine the minimum residual range according to the second correspondence The corresponding second index value, the second index value is used for transmission to the decoding end.

Optionally, when the processor determines the first target number of bits occupied by the first residual after compression, according to the corresponding pixel value in the minimum block to which the first pixel value belongs. The maximum residual of the residual, the initial allocation of the number of bits is performed on the first residual to determine the first target number of bits.

Optionally, the processor is further configured to, if there is a remaining number of bits after the initial allocation, preferentially assign the remaining number of bits to the corresponding pixel values of the first data to be encoded at the division boundary. of the residuals.

Optionally, the processor is further configured to, if the compressed residual is larger than the uncompressed residual, modify the value of the compressed residual to the value of the uncompressed residual.

Optionally, when determining the residual corresponding to each pixel value in the plurality of pixel values according to the first minimum pixel value, the processor is configured to use the compressed first minimum pixel value. Reconstruction is performed to obtain the reconstructed first minimum pixel value; the residual between each of the pixel values and the reconstructed first minimum pixel value is determined as the residual corresponding to each pixel value.

Optionally, the processor is configured to, when compressing the first minimum pixel value, determine the largest residual error among the residual errors corresponding to each pixel value in the plurality of pixel values; The maximum residual determines the second target number of bits occupied by the first minimum pixel value after compression; the first minimum pixel value is compressed according to the second target number of bits.

Optionally, the first data to be encoded is obtained by dividing a first data block to be encoded of the image to be encoded.

Optionally, the to-be-encoded image includes an original raw image.

Optionally, the first data block to be encoded is obtained based on the encoded data block corresponding to the first color component of the image to be encoded.

Optionally, the encoded data block corresponding to the first color component is obtained by de-interleaving the original encoded data block of the image to be encoded.

Optionally, the first data block to be encoded is obtained by compressing the encoded data block corresponding to the first color component.

Optionally, the processor is configured to compress each pixel value in the encoded data block corresponding to the first color component by logarithmic transformation when compressing the encoded data block corresponding to the first color component. .

For the various implementation manners of the image encoding apparatus provided above, reference may be made to the relevant descriptions above for the specific implementation, which will not be repeated here.

In the image encoding device provided by the embodiment of the present application, the encoding end can determine the residual corresponding to each pixel value according to the first minimum pixel value corresponding to each pixel value, and can compress the residual corresponding to each pixel value and the first minimum pixel value. Then, it is transmitted to the decoding end, so that the decoding end can reconstruct and obtain each pixel value based on the residual corresponding to each pixel value and the first minimum pixel value. It can be seen that, because only the first minimum pixel value and the residual corresponding to each pixel value need to be transmitted between the encoding end and the decoding end, and the original pixel value of each pixel does not need to be transmitted, the amount of data to be transmitted is reduced. In addition, since the residual corresponding to each pixel value is determined based on the first minimum pixel value, the pixel value (dark part of the image) close to the first minimum pixel value can retain high precision, which is consistent with the human eye's sensitivity to changes in brightness. The perceptual sensitivity is consistent with the law that the dark part is higher than the bright part, which improves the compression performance while ensuring good image quality.

Referring to FIG. 8 below, FIG. 8 is a schematic structural diagram of an image encoding apparatus provided by an embodiment of the present application. The apparatus comprises: a processor 810 and a memory 820 storing a computer program, the processor implements the following steps when executing the computer program:

Optionally, the first data to be decoded includes the plurality of compressed residuals, and the processor is configured to, when decompressing the first data to be decoded, decompress the plurality of compressed residuals. The residuals are decompressed by inverse log transformation.

Optionally, when the processor decompresses the compressed residuals through inverse logarithmic transformation, the processor is configured to, according to the minimum residual range to which the first residual belongs, and the first residual At the first target number of bits occupied after compression, decompress the compressed first residual, wherein the first residual is the residual corresponding to the first pixel value, and the first pixel value is any pixel value in the block.

Optionally, according to the minimum residual range to which the first residual belongs, and the first target number of bits occupied by the first residual after compression, the processor performs a process on the compressed first residual. When performing decompression, it is used to decompress the compressed first residual through inverse gamma transform according to the first target number of bits, so as to obtain that the first residual is within the minimum residual range The corresponding ratio; the first residual is determined according to the ratio and the minimum residual range.

Optionally, the minimum residual error range to which the first residual belongs is determined according to a second correspondence relationship and a second index value, and the second correspondence relationship is a pre-configured minimum residual error range and the second index value. Corresponding relationship, the second index value is obtained by decoding from the first data to be decoded.

Optionally, when determining the minimum residual error range to which the first residual belongs, the processor is configured to: determine the minimum residual error range according to the second correspondence and the second index value; according to the first correspondence and The first index value determines the adjustment amount corresponding to the minimum residual error range; the minimum residual error range is adjusted according to the adjustment amount, and the adjusted minimum residual error range is determined as the one to which the first residual error belongs. Minimum residual range;

Wherein, the first correspondence is the correspondence between the pre-configured adjustment amount and the first index value, the second correspondence is the correspondence between the pre-configured minimum residual error range and the second index value, the first Both the index value and the second index value are decoded from the first data to be decoded.

Optionally, when the processor determines, according to the first minimum pixel value and the plurality of residuals, the plurality of pixel values for the plurality of pixels of the to-be-decoded image, the processor is configured to, according to the first The minimum pixel value and the multiple residuals are decompressed to obtain multiple pixel values of multiple pixels of the to-be-decoded image.

Optionally, the first data to be decoded includes the compressed first minimum pixel value, and the processor is used to decompress the compressed first data to be decoded. A minimum pixel value is decompressed by an inverse log transform.

For the various implementation manners of the image decoding apparatus provided above, reference may be made to the relevant descriptions above for the specific implementation, which will not be repeated here.

In the image decoding apparatus provided in the embodiment of the present application, the decoding end can decompress the first data to be decoded to obtain the first minimum pixel value and multiple residuals, and can, according to the multiple residuals and the first minimum pixel value, Multiple pixel values of the resulting image can be reconstructed. It can be seen that, because only the first minimum pixel value and the residual corresponding to each pixel value need to be transmitted between the encoding end and the decoding end, and the original pixel value of each pixel does not need to be transmitted, the amount of data to be transmitted is reduced. Moreover, each pixel value is reconstructed based on the first minimum pixel value, so the pixel value (dark part of the image) close to the first minimum pixel value can retain high precision, which is related to the human eye's perception sensitivity to brightness changes. The dark part is higher than the bright part in accordance with the law, which can improve the compression performance and ensure good image quality.

The embodiments of the present application further provide a computer-readable storage medium, where the computer-readable storage medium stores a computer program, and when the computer program is executed by a processor, implements the image encoding method provided by the embodiments of the present application.

The embodiment of the present application further provides a computer-readable storage medium, where the computer-readable storage medium stores a computer program, and the computer program implements the image decoding method provided by the embodiment of the present application when the computer program is executed by the processor.

Various implementations are provided above for each protection subject. On the basis of no conflict or contradiction, those skilled in the art can freely combine various implementations according to the actual situation, thereby forming various technical solutions. . However, this application document is limited in space and cannot describe all the technical solutions obtained by combination, but it can be understood that these technical solutions that cannot be developed also belong to the scope disclosed in the embodiments of this application.

Embodiments of the present application may take the form of a computer program product implemented on one or more storage media (including, but not limited to, disk storage, CD-ROM, optical storage, etc.) having program code embodied therein. Computer-usable storage media includes permanent and non-permanent, removable and non-removable media, and storage of information can be accomplished by any method or technology. Information may be computer readable instructions, data structures, modules of programs, or other data. Examples of computer storage media include, but are not limited to, phase-change memory (PRAM), static random access memory (SRAM), dynamic random access memory (DRAM), other types of random access memory (RAM), read only memory (ROM), Electrically Erasable Programmable Read Only Memory (EEPROM), Flash Memory or other memory technology, Compact Disc Read Only Memory (CD-ROM), Digital Versatile Disc (DVD) or other optical storage, Magnetic tape cassettes, magnetic tape magnetic disk storage or other magnetic storage devices or any other non-transmission medium that can be used to store information that can be accessed by a computing device.

It should be noted that, in this document, relational terms such as first and second are only used to distinguish one entity or operation from another entity or operation, and do not necessarily require or imply any relationship between these entities or operations. any such actual relationship or sequence exists. The terms "comprising", "comprising" or any other variation thereof are intended to encompass non-exclusive inclusion such that a process, method, article or device comprising a list of elements includes not only those elements, but also other not expressly listed elements, or also include elements inherent to such a process, method, article or apparatus. Without further limitation, an element qualified by the phrase "comprising a..." does not preclude the presence of additional identical elements in a process, method, article or apparatus that includes the element.

The methods and devices provided by the embodiments of the present application have been introduced in detail above, and specific examples are used to illustrate the principles and implementations of the present application. At the same time, for those of ordinary skill in the art, according to the idea of the application, there will be changes in the specific implementation and application scope. In summary, the content of this specification should not be construed as a limitation to the application. .

Claims

An image coding method, comprising:

acquiring first data to be encoded corresponding to the image to be encoded, the first data to be encoded includes multiple pixel values of multiple pixels of a block in the image to be encoded;

determining a first minimum pixel value for the plurality of pixel values from the plurality of pixel values;

determining, according to the first minimum pixel value, a residual corresponding to each pixel value in the plurality of pixel values;

The residual corresponding to each pixel value and the first minimum pixel value are compressed.
The method according to claim 1, wherein the compression ratio of the residual corresponding to the pixel value is positively correlated with the size of the residual corresponding to the pixel value.
The method according to claim 2, wherein compressing the residual corresponding to each pixel value comprises:

The residual corresponding to each pixel value is compressed by logarithmic transformation.
The method according to claim 3, wherein the compressing the residual corresponding to each pixel value by logarithmic transformation, comprising:

determining the minimum residual range to which the first residual belongs, wherein the first residual is the residual corresponding to the first pixel value, and the first pixel value is any pixel value in the block;

determining the first target number of bits occupied by the first residual after compression;

The first residual is compressed according to the minimum residual range and the first target number of bits.
The method according to claim 4, wherein the compressing the first residual according to the minimum residual range and the first target number of bits comprises:

converting the first residual into the proportion of the minimum residual range;

The ratio is compressed into a numerical range corresponding to the first target number of bits by gamma transformation.
The method according to claim 4, wherein the minimum residual range to which the first residual belongs is the minimum residual range corresponding to the block, and the minimum residual range corresponding to the block is in the following manner Sure:

determining the maximum residual of the residual corresponding to each pixel value in the block;

The minimum residual error range is determined according to the maximum residual error.
The method of claim 4, wherein the block is divided into a plurality of sub-blocks, the first pixel value belongs to a first sub-block of the plurality of sub-blocks, the first The minimum residual range to which the residual belongs is the minimum residual range corresponding to the first sub-block, and the minimum residual range corresponding to the first sub-block is determined in the following manner:

determining the maximum residual of the residuals corresponding to each pixel value in the first subblock;

The minimum residual error range is determined according to the maximum residual error.
The method according to claim 7, wherein the method further comprises:

determining the second minimum pixel value of the plurality of pixel values in the first subblock;

According to the minimum value difference between the second minimum pixel value and the first minimum pixel value, the minimum residual error range corresponding to the first sub-block is adjusted, so as to determine the relative value of the first sub-block The adjusted minimum residual range for the block.
The method according to claim 8, wherein, according to the minimum value difference between the second minimum pixel value and the first minimum pixel value, for the minimum value corresponding to the first sub-block The residual error range is adjusted to determine the adjusted residual error range for the first sub-block, including:

The adjustment amount corresponding to the minimum difference value is determined according to the first correspondence, and the minimum residual range corresponding to the first sub-block is adjusted according to the adjustment amount, and the first correspondence is pre-configured The corresponding relationship between the minimum difference value and the adjustment amount.
The method according to claim 9, wherein the first correspondence is the correspondence between the minimum difference value, the adjustment amount and the first index value, and the method further comprises:

A first index value corresponding to the adjustment amount is determined according to the first correspondence, and the first index value is used for transmission to the decoding end.
The method according to claim 6 or 7, wherein the determining the minimum residual error range according to the maximum residual error comprises:

A minimum residual error range corresponding to the maximum residual error is determined according to a second correspondence relationship, where the second correspondence relationship is a preconfigured correspondence relationship between the maximum residual error and the minimum residual error range.
The method according to claim 11, wherein the second correspondence is a correspondence between the maximum residual error, the minimum residual error range and the second index value, and the method further comprises:

A second index value corresponding to the minimum residual error range is determined according to the second correspondence, and the second index value is used for transmission to the decoding end.
The method according to claim 4, wherein the determining the first target number of bits occupied by the first residual after compression comprises:

According to the maximum residual of the residuals corresponding to each pixel value in the smallest block to which the first pixel value belongs, an initial allocation of the number of bits is performed on the first residual to determine the first target bit. number.
The method of claim 13, wherein the method further comprises:

If there is a remaining number of bits after the initial allocation, the remaining number of bits is preferentially allocated to the residual corresponding to the pixel value of the first data to be encoded at the division boundary.
The method according to claim 3, wherein the method further comprises:

If the compressed residual is larger than the uncompressed residual, the value of the compressed residual is modified to the value of the uncompressed residual.
The method according to claim 1, wherein the determining, according to the first minimum pixel value, a residual corresponding to each pixel value in the plurality of pixel values comprises:

Use the compressed first minimum pixel value for reconstruction to obtain the reconstructed first minimum pixel value;

A residual between each of the pixel values and the reconstructed first minimum pixel value is determined as the residual corresponding to each pixel value.
The method according to claim 1, wherein compressing the first minimum pixel value comprises:

determining the largest residual among the residuals corresponding to each of the plurality of pixel values;

determining the second target number of bits occupied by the first minimum pixel value after compression according to the maximum residual error;

The first minimum pixel value is compressed according to the second target number of bits.
The method according to claim 1, wherein the first data to be encoded is obtained by dividing a first data block to be encoded of the image to be encoded.
The method of claim 18, wherein the to-be-encoded image comprises an original raw image.
The method according to claim 19, wherein the first data block to be encoded is obtained based on the encoded data block corresponding to the first color component of the image to be encoded.
The method according to claim 20, wherein the encoded data block corresponding to the first color component is obtained by de-interleaving the original encoded data block of the image to be encoded.
The method according to claim 20, wherein the first data block to be encoded is obtained by compressing the encoded data block corresponding to the first color component.
The method according to claim 21, wherein compressing the encoded data block corresponding to the first color component comprises:

Each pixel value in the encoded data block corresponding to the first color component is compressed by logarithmic transformation.
An image decoding method, characterized in that the method comprises:

obtaining the first to-be-decoded data corresponding to the to-be-decoded image;

decompressing the first data to be decoded to determine a first minimum pixel value and a plurality of residuals;

From the first minimum pixel value and the plurality of residuals, a plurality of pixel values for a plurality of pixels of the decoded image are determined.
The method according to claim 24, wherein the compression ratio of the residual corresponding to the pixel value is positively correlated with the size of the residual corresponding to the pixel value.
The method according to claim 25, wherein the first data to be decoded comprises the plurality of compressed residuals, and the decompressing the first data to be decoded comprises:

The compressed residuals are decompressed by inverse log transformation.
The method according to claim 26, wherein the decompressing the compressed residuals by inverse logarithmic transformation comprises:

The compressed first residual is decompressed according to the minimum residual range to which the first residual belongs and the first target number of bits occupied by the first residual after compression, wherein the first residual is decompressed. A residual is the residual corresponding to the first pixel value, and the first pixel value is any pixel value in the block.
The method according to claim 27, wherein, according to the minimum residual range to which the first residual belongs and the first target number of bits occupied by the first residual after compression, the compressed The first residual is decompressed, including:

According to the first target number of bits, the compressed first residual is decompressed by inverse gamma transform to obtain the proportion of the first residual corresponding to the minimum residual range;

The first residual is determined according to the proportion and the minimum residual range.
The method according to claim 27, wherein the minimum residual error range to which the first residual belongs is determined according to a second correspondence relationship and a second index value, and the second correspondence relationship is a pre-configured minimum residual error The correspondence between the residual range and the second index value, where the second index value is decoded from the first data to be decoded.
The method according to claim 27, wherein the minimum residual error range to which the first residual belongs is determined in the following manner:

determining the minimum residual error range according to the second correspondence and the second index value;

Determine the adjustment amount corresponding to the minimum residual error range according to the first correspondence and the first index value;

Adjust the minimum residual error range according to the adjustment amount, and determine the adjusted minimum residual error range as the minimum residual error range to which the first residual belongs;

Wherein, the first correspondence is the correspondence between the pre-configured adjustment amount and the first index value, the second correspondence is the correspondence between the pre-configured minimum residual error range and the second index value, the first Both the index value and the second index value are decoded from the first data to be decoded.
The method according to claim 24, wherein the determining a plurality of pixel values for a plurality of pixels of the to-be-decoded image according to the first minimum pixel value and the plurality of residuals comprises:

Decompressing is performed according to the first minimum pixel value and the plurality of residuals to obtain a plurality of pixel values of a plurality of pixels of the image to be decoded.
The method according to claim 24, wherein the first data to be decoded comprises the compressed first minimum pixel value, and the decompressing the first data to be decoded comprises:

The compressed first minimum pixel value is decompressed by inverse logarithmic transformation.
An image encoding device, characterized in that it comprises: a processor and a memory storing a computer program, wherein the processor implements the following steps when executing the computer program:

acquiring first data to be encoded corresponding to the image to be encoded, the first data to be encoded includes multiple pixel values of multiple pixels of a block in the image to be encoded;

determining a first minimum pixel value for the plurality of pixel values from the plurality of pixel values;

determining, according to the first minimum pixel value, a residual corresponding to each pixel value in the plurality of pixel values;

The residual corresponding to each pixel value and the first minimum pixel value are compressed.
The apparatus according to claim 33, wherein the compression ratio of the residual corresponding to the pixel value is positively correlated with the size of the residual corresponding to the pixel value.
The device according to claim 34, wherein the processor is configured to compress the residual corresponding to each pixel value through logarithmic transformation when compressing the residual corresponding to each pixel value .
The device according to claim 35, wherein the processor is configured to determine the minimum residual error range to which the first residual belongs when compressing the residual corresponding to each pixel value by logarithmic transformation, Wherein, the first residual is the residual corresponding to the first pixel value, and the first pixel value is any pixel value in the block; determine the first residual occupied by the first residual after compression target number of bits; compressing the first residual according to the minimum residual range and the first target number of bits.
The apparatus according to claim 36, wherein when the processor compresses the first residual according to the minimum residual range and the first target number of bits, the processor is configured to: The first residual is converted into a proportion of the minimum residual error range; the proportion is compressed into a numerical range corresponding to the first target number of bits through gamma transformation.
The apparatus according to claim 36, wherein the minimum residual range to which the first residual belongs is a minimum residual range corresponding to the block, and the processor is determining the minimum residual range corresponding to the block. When the residual error range is used, the maximum residual error of the residual error corresponding to each pixel value in the block is determined; the minimum residual error range is determined according to the maximum residual error.
The apparatus of claim 36, wherein the block is divided into a plurality of sub-blocks, the first pixel value belongs to a first sub-block of the plurality of sub-blocks, and the first pixel value belongs to a first sub-block of the plurality of sub-blocks. The minimum residual range to which the residual belongs is the minimum residual range corresponding to the first sub-block, and the processor is used for determining the first sub-block when determining the minimum residual range corresponding to the first sub-block. The maximum residual of the residual corresponding to each pixel value in a sub-block; the minimum residual range is determined according to the maximum residual.
The device according to claim 39, wherein the processor is further configured to determine a second minimum pixel value of a plurality of pixel values in the first sub-block; according to the second minimum pixel value and the minimum value difference between the first minimum pixel value, the minimum residual error range corresponding to the first sub-block is adjusted to determine the adjusted minimum residual error about the first sub-block scope.
34. The apparatus of claim 34, wherein the processor performs the first sub-block according to a minimum value difference between the second minimum pixel value and the first minimum pixel value The corresponding minimum residual error range is adjusted to determine the adjusted residual error range about the first sub-block, and is used to determine the adjustment amount corresponding to the minimum value difference according to the first correspondence, and according to the The adjustment amount adjusts the minimum residual error range corresponding to the first sub-block, and the first correspondence relationship is a correspondence relationship between a pre-configured minimum difference value and an adjustment amount.
The apparatus according to claim 41, wherein the first correspondence is a correspondence between the minimum value difference, the adjustment amount and the first index value, and the processor is further configured to: according to the first correspondence The relationship determines the first index value corresponding to the adjustment amount, and the first index value is used for transmission to the decoding end.
The apparatus according to claim 38 or 39, wherein, when the processor determines the minimum residual error range according to the maximum residual error, the processor is configured to determine, according to a second correspondence relationship, a Minimum residual error range, the second correspondence is a pre-configured correspondence relationship between the maximum residual error and the minimum residual error range.
The apparatus according to claim 43, wherein the second correspondence is a correspondence between a maximum residual error, a minimum residual error range and a second index value, and the processor is further configured to: The correspondence determines the second index value corresponding to the minimum residual error range, and the second index value is used for transmission to the decoding end.
The apparatus according to claim 36, wherein when the processor determines the first target number of bits occupied by the first residual after compression, the processor is configured to: The maximum residual of the residuals corresponding to each pixel value in the small block, and initial allocation of the number of bits is performed on the first residual to determine the first target number of bits.
The device according to claim 45, wherein the processor is further configured to, if there is a remaining number of bits after the initial allocation, preferentially allocate the remaining number of bits to the first to-be-coded The residual corresponding to the pixel value of the data at the partition boundary.
The apparatus according to claim 35, wherein the processor is further configured to, if the compressed residual is larger than the uncompressed residual, modify the value of the compressed residual to the uncompressed residual value of .
34. The apparatus of claim 33, wherein the processor is configured to utilize compression when determining the residual corresponding to each pixel value in the plurality of pixel values according to the first minimum pixel value. The reconstructed first minimum pixel value is reconstructed to obtain the reconstructed first minimum pixel value; the residual between each of the pixel values and the reconstructed first minimum pixel value is determined as the corresponding pixel value. the residuals.
The apparatus according to claim 33, wherein the processor is configured to, when compressing the first minimum pixel value, determine the residual value corresponding to each pixel value in the plurality of pixel values The maximum residual in the difference; determine the second target number of bits occupied by the first minimum pixel value after compression according to the maximum residual; to compress.
The apparatus according to claim 33, wherein the first data to be encoded is obtained by dividing a first data block to be encoded of the image to be encoded.
The apparatus of claim 50, wherein the to-be-encoded image comprises an original raw image.
The apparatus according to claim 51, wherein the first data block to be encoded is obtained based on the encoded data block corresponding to the first color component of the image to be encoded.
The apparatus according to claim 52, wherein the encoded data block corresponding to the first color component is obtained by de-interleaving the original encoded data block of the image to be encoded.
The apparatus according to claim 52, wherein the first data block to be encoded is obtained by compressing the encoded data block corresponding to the first color component.
The apparatus according to claim 53, wherein when the processor compresses the encoded data block corresponding to the first color component, the processor is configured to: compress the encoded data block corresponding to the first color component Each pixel value is compressed by logarithmic transformation.
An image decoding device, characterized in that it comprises: a processor and a memory storing a computer program, wherein the processor implements the following steps when executing the computer program:

obtaining the first to-be-decoded data corresponding to the to-be-decoded image;

decompressing the first data to be decoded to determine a first minimum pixel value and a plurality of residuals;

From the first minimum pixel value and the plurality of residuals, a plurality of pixel values for a plurality of pixels of the decoded image are determined.
The apparatus according to claim 56, wherein the compression ratio of the residual corresponding to the pixel value is positively correlated with the size of the residual corresponding to the pixel value.
The apparatus according to claim 57, wherein the first data to be decoded comprises the plurality of compressed residuals, and the processor is used for decompressing the first data to be decoded for decompressing the first data to be decoded. , and decompress the compressed residuals through inverse logarithmic transformation.
The apparatus according to claim 58, wherein when the processor decompresses the compressed residuals through inverse logarithmic transformation, the processor is configured to: according to the smallest residual to which the first residual belongs range, and the first target number of bits occupied by the first residual after compression, and decompress the compressed first residual, where the first residual is the value corresponding to the first pixel value Residual, the first pixel value is any pixel value in the block.
The apparatus according to claim 59, wherein, according to the minimum residual range to which the first residual belongs and the first target number of bits occupied by the first residual after compression, the processor determines the When the compressed first residual is decompressed, it is used for decompressing the compressed first residual through inverse gamma transform according to the first target number of bits to obtain the first residual. The proportion of the residual error corresponding to the minimum residual error range; the first residual error is determined according to the proportion and the minimum residual error range.
The apparatus according to claim 59, wherein the minimum residual error range to which the first residual belongs is determined according to a second correspondence relationship and a second index value, and the second correspondence relationship is a pre-configured minimum residual error The correspondence between the residual range and the second index value, where the second index value is decoded from the first data to be decoded.
The apparatus according to claim 59, wherein when determining the minimum residual range to which the first residual belongs, the processor is configured to determine the minimum residual according to a second correspondence and a second index value difference range; determine the adjustment amount corresponding to the minimum residual error range according to the first correspondence and the first index value; adjust the minimum residual error range according to the adjustment amount, and determine the adjusted minimum residual error range is the minimum residual range to which the first residual belongs;

Wherein, the first correspondence is the correspondence between the pre-configured adjustment amount and the first index value, the second correspondence is the correspondence between the pre-configured minimum residual error range and the second index value, the first Both the index value and the second index value are decoded from the first data to be decoded.
The apparatus according to claim 56, wherein the processor determines a plurality of pixel values for a plurality of pixels of the to-be-decoded image according to the first minimum pixel value and the plurality of residuals is used to perform decompression according to the first minimum pixel value and the multiple residuals to obtain multiple pixel values of multiple pixels of the to-be-decoded image.
The apparatus according to claim 56, wherein the first data to be decoded comprises the compressed first minimum pixel value, and the processor decompresses the first data to be decoded by using Then, the compressed first minimum pixel value is decompressed by inverse logarithmic transformation.
A computer-readable storage medium, characterized in that, the computer-readable storage medium stores a computer program, and when the computer program is executed by a processor, the image encoding method according to any one of claims 1 to 23 is implemented.
A computer-readable storage medium, characterized in that, the computer-readable storage medium stores a computer program, and when the computer program is executed by a processor, the image decoding method according to any one of claims 24 to 32 is implemented.