Lossless compression method for eliminating inter-frame redundancy based on blocks
By segmenting video frames and constructing difference frames, combined with entropy coding, the problem of low efficiency in eliminating inter-frame redundancy in lossless video compression is solved, achieving efficient and robust lossless video compression, suitable for scenarios such as medical imaging and professional video archiving.
Patent Information
- Authority / Receiving Office
- CN · China
- Patent Type
- Applications(China)
- Current Assignee / Owner
- CHANGZHOU WHISPER TECH CO LTD
- Filing Date
- 2026-02-09
- Publication Date
- 2026-04-21
AI Technical Summary
Existing lossless video compression methods are inefficient in inter-frame prediction, easily introducing irreversible changes and prediction errors, leading to a decline in video quality. They are particularly unable to meet the requirements of pixel-level lossless reconstruction in scenarios such as medical imaging and professional video archiving.
A block-based inter-frame redundancy elimination method is adopted. By dividing the video frame into blocks, the pixel difference between the current frame and the previous frame is used to divide the candidate blocks and match motion vectors. The difference frame is constructed and divided into coarse and fine granular parts. Sub-blocks in different states are recorded and compression is performed in combination with entropy coding.
It improves the efficiency and robustness of lossless video compression, adapts to video content with different resolutions and motion complexity, maintains high compression efficiency in static areas and stable prediction performance in complex motion areas, and achieves integer reversibility and high compression efficiency.
Smart Images
Figure CN121908016A_ABST
Abstract
Description
Technical Field
[0001] This invention relates to the field of lossless compression technology, and in particular to a lossless compression method based on block-based elimination of inter-frame redundancy. Background Technology
[0002] There is a large amount of temporal redundancy between adjacent frames in a video sequence: that is, stationary or slowly moving objects in a scene exhibit highly similar pixel distributions in consecutive frames. Traditional video coding utilizes this temporal redundancy during inter-frame prediction (motion estimation and motion compensation), which significantly reduces the bitrate in lossy coding scenarios, resulting in unclear videos, the potential elimination of key pixels, and the loss of crucial information. However, for lossless compression scenarios that require precise pixel reconstruction (such as medical image preservation, professional video archiving, and post-production material transmission), conventional lossy prediction and quantization steps are unacceptable, and redundancy must be removed without losing any original information.
[0003] However, current color coding methods for continuous vision (H.264) mostly employ prediction and interpolation, which exposes several drawbacks in lossless inter-frame compression: 1. Commonly used sub-pixel interpolation or approximate prediction introduces irreversible changes, making it difficult to meet pixel-level lossless reconstruction requirements; 2. Prediction errors propagate along the prediction chain, which may lead to significant quality deviations over long-term encoding. 3. When the prediction accuracy is insufficient, the residual distribution is scattered, and the entropy coding efficiency decreases. Summary of the Invention
[0004] To address the shortcomings of existing methods, this invention solves the problem of low efficiency in lossless inter-frame compression.
[0005] The technical solution adopted in this invention is: a lossless compression method based on block-based elimination of inter-frame redundancy, comprising the following steps: Step 1: Acquire the video of a specific shot and segment the video into frames; Step 2: Name adjacent frames and divide them into sub-grids; Step 3: Divide candidate blocks using the absolute error of the pixel difference between corresponding sub-blocks in the current frame and the previous frame; As a preferred embodiment of the present invention, step three specifically includes: Divide the current frame and the previous frame into several sub-blocks according to the sub-grid size; For a sub-block in the current frame, the SAD value is calculated block by block by sliding window after extending the sub-block at the same position in the previous frame by a preset number of pixels. If the SAD values are the same, they are set as the same candidate block; otherwise, the block with the smallest SAD value is selected as the similar candidate block. Output the motion vector MV and state vector of the same or similar candidate blocks.
[0006] Step 4: Construct candidate frames using candidate blocks; Step 5: Subtract the current frame from the candidate frame to obtain the difference frame; perform coarse and fine-grained sub-grid division on the difference frame; output the recording status of the sub-blocks in the sub-grid after granular division; In a preferred embodiment of the present invention, step five specifically includes: Step 51: Perform coarse-grained grid difference frame division; record state division based on pixel difference of sub-blocks; Step 511: When the pixel difference of all pixels in a sub-block is 0, record it as a copy state; Step 512: When the pixel difference value of all pixels in a sub-block is the same, record it as a monochrome state; Step 513: When the pixel difference of all pixels in a sub-block is the same as the pixel difference of a neighboring block, it is recorded as a neighborhood consistency. If any of steps 511 to 513 is not satisfied, perform the fine-grained mesh difference frame division in step 52.
[0007] In a preferred embodiment of the present invention, step 52 specifically includes: Step 521: Divide the coarse-grained mesh into a fine-grained mesh of size 2*2, and set the step size to 2; Step 522: When the pixel difference value of all pixels in a sub-block is 0, record it as a copy state; Step 523: When the pixel difference value of all pixels in a sub-block is the same, record it as a monochrome state; Step 524: When the pixel difference of all pixels in a sub-block is the same as the pixel difference of a neighboring block, it is recorded as a neighborhood consistency state. Step 525: When there are two possible pixel differences among all pixels in a sub-block, record it as a double constant state; Step 526: When there are three types of pixel differences among all pixels in a certain sub-block, it is recorded as a three-constant state; Step 527: When there are four types of pixel differences among all pixels in a certain sub-block, it is recorded as a four-constant state.
[0008] Step 6: Divide the current frame into coarse and fine-grained sub-mesh, and output the recording status of the sub-blocks in the sub-mesh after granular division; In a preferred embodiment of the present invention, step six specifically includes: Step 611: When the pixel difference value of all pixels in a sub-block is the same, record it as a monochrome state; Step 612: When the pixel difference of all pixels in a sub-block is the same as the pixel difference of a neighboring block, it is recorded as a neighborhood consistency state. If either step 611 or step 612 is not satisfied, perform the fine mesh difference frame division in step 62.
[0009] In a preferred embodiment of the present invention, step 62 specifically includes: Step 621: Divide the coarse-grained mesh into a fine-grained mesh of size 2*2, and set the step size to 2; Step 622: When the pixel difference value of all pixels in a sub-block is the same, record it as a monochrome state; Step 623: When the pixel difference of all pixels in a sub-block is the same as the pixel difference of a neighboring block, it is recorded as neighborhood consistency. Step 624: When there are two possible pixel differences among all pixels in a sub-block, record it as a double constant state; Step 625: When there are three types of pixel differences among all pixels in a certain sub-block, record it as a three-constant state; Step 626: When there are four types of pixel differences among all pixels in a certain sub-block, it is recorded as a four-constant state.
[0010] Step 7: Set different priorities for the recording status of the current frame and the difference frame, extract the recording status from the sub-blocks of the current frame and the difference frame in order of priority, and output the new frame. In a preferred embodiment of the present invention, the priority order is customized according to the different recorded states.
[0011] In a preferred embodiment of the present invention, the container corresponding to the new frame is output, and the container is compressed using entropy coding.
[0012] As a preferred embodiment of the present invention, a lossless compression system based on block-based elimination of inter-frame redundancy includes: a memory for storing instructions executable by a processor; and a processor for executing the instructions to implement a lossless compression method based on block-based elimination of inter-frame redundancy.
[0013] As a preferred embodiment of the present invention, a computer-readable medium storing computer program code implements a lossless compression method based on block-based elimination of inter-frame redundancy when executed by a processor.
[0014] The beneficial effects of this invention are: 1. The block-based strategy can adapt to video content with different resolutions and motion complexities. It can achieve higher compression efficiency in static and weak motion areas, and can still maintain stable prediction performance in complex motion areas, with stronger overall robustness. 2. For lossless scenarios, it is necessary to design a block-based inter-frame compression scheme that can guarantee integer and reversibility and remain robust under rapid motion and scene switching. 3. The block-based inter-frame redundancy elimination method divides the frame into several small blocks, searches for the best matching region in the reference frame for each block, describes the displacement through motion vectors, and encodes the residual between the current block and the reference block. 4. For lossless targets, this process must be reversible, without introducing lost quantization or approximate interpolation, and the transformation should use an integer reversible method. Since the residual has a low amplitude and high statistical concentration, it can be further compressed in subsequent entropy coding, thereby improving image compression efficiency and transmission speed. Attached Figure Description
[0015] Figure 1 This is a flowchart of the lossless compression method based on block-based elimination of inter-frame redundancy of the present invention; Figure 2 This is a schematic diagram of the first subgrid of image B in this invention; Figure 3 This is a schematic diagram of the first sub-grid search range of image A and image B in this invention; Figure 4 This refers to the relative displacement of the first subgrid between image A and image B in this invention; Figure 5 This is a schematic diagram of the two constants of the present invention; Figure 6 This is a schematic diagram of the three constants of the present invention; Figure 7 This is a schematic diagram of the four constants of the present invention; Figure 8 This is the re-extracted state marker of the present invention. Detailed Implementation
[0016] The present invention will be further described below with reference to the accompanying drawings and embodiments. The drawings are simplified schematic diagrams, which only illustrate the basic structure of the present invention in a schematic manner, and therefore only show the components related to the present invention.
[0017] like Figure 1 As shown, a lossless compression method based on block-based elimination of inter-frame redundancy includes the following steps: Step 1: Acquire the video of a specific shot and segment the video into frames; The first frame is losslessly compressed using conventional intra-frame compression methods, and the compressed file of the first frame is saved; inter-frame compression of the present invention is performed starting from the second frame. Step 2: Name and divide adjacent frames into blocks; If the second frame is taken as the previous frame A, the third frame as the current frame B; and so on, if the third frame is taken as the previous frame A, the fourth frame as the current frame B; The following explanation uses frames 2 and 3 as examples; Divide the current frame B into sub-grids; taking an 8*8 pixel grid as an example, process the current frame B block by block; the grid size can be customized; for example, m*n, where m and n are determined according to the frame image size. If the resolution is large, the sub-blocks can be divided larger, usually not exceeding 256*256. If the resolution is low, the sub-grids can be divided smaller, with a minimum of 2*2. Step 3: Based on the candidate search regions of the current frame and the previous frame; Sub-block B of the current frame i The corresponding position in the previous frame A is divided into sub-blocks A of the previous frame. i B i In A i Perform a sliding window search within the corresponding position ±k pixels according to the sub-grid size to find the match with B. i The search stops when sub-blocks have the same pixel value; when a similar pixel value is found, A is set to... i Add to the candidate search region set; like Figure 2 As shown, the green dot represents an 8x8 B. i Sub-block, with coordinates (y, x); Figure 3 The yellow area is Figure 2 B i Sub-block in A i Extend the range by ±32 pixels in the four directions corresponding to the position; that is, (y±32, x±32). Figure 2 When a matching or similar candidate block is found within the yellow area, its displacement coordinates are recorded; for example... Figure 4 As shown, green represents the sub-block position in frame B, and orange represents the candidate block matched in frame A; Candidate blocks are calculated by sum of absolute differences (SAD), which is given by the formula:
[0018] Where A(i, j) is the pixel intensity of block A at position (i, j), and B(i, j) is the pixel intensity of block B at the corresponding position.
[0019] Candidate blocks with the same SAD value are 0, and similar candidate blocks are the blocks with the smallest SAD value, for example: B i The sub-block is 8*8, A i The sub-block is (8+32+32,8+32+32)=(72,72), so move A... iThe first 8*8 block in the sub-block is slid windowed with a step size of 1 to obtain the second 8*8 block, and so on, until the last block at position (72,72) is obtained. Calculate the SAD value for each block. Based on the coordinates (i.e., motion vectors MV) of the same or similar candidate blocks in A, obtain the same or similar motion vectors MV of B. The formula for the relative displacement of a B-frame sub-block within frame A, i.e., the motion vector MV, is: MV_Y = ym MV_X = xn Where (m,n) are the position coordinates of the candidate block in the sub-block of the previous frame A; Output the motion vector of frame B and the corresponding candidate block state vector; that is, B1...B i =(MV1,1;MV2,0;MV3,1;...MV i ,1;), where MV1,1 represents the motion vector of B1, 1 indicates a similar candidate block, and MV2,0 represents the motion vector of B2, 0 indicates a similar candidate block; B i There is only one state: 0 or 1.
[0020] Step 4: Construct candidate frame A using candidate blocks in frame A ’ ; That is, A ’ Constructed from similar or identical candidate blocks in A, with the same size as A; Step 5: Based on the current frame B and candidate frame A ’ Construct the difference frame C; C=BA ’ That is, each pixel in frame B is compared with that in frame A. ’ Subtract each pixel from the others; Divide C into coarse-grained and fine-grained subgrids, with the first subgrid being the coarse-grained subgrid and the second subgrid being the fine-grained subgrid. Typically, the first subgrid is larger than the second subgrid, and the first subgrid is an integer multiple of the second subgrid. The size of the first subgrid is the same as the subgrid size in step two, while the second subgrid is the smallest granularity. For example, the first subgrid size is set to 8*8, and the second subgrid size is set to 2*2; the first subgrid is four times the size of the second subgrid. Step 51: Perform the first sub-grid determination for frame C: Step 511: When the pixel difference of all pixels in a sub-block of C is 0, record the "copy" flag as 0; For example, the RGB values of all 64 pixels in this sub-block are 000; Step 512: When the pixel difference value of all pixels in a certain sub-block is the same, record it as "monochrome" and mark it as 1; that is, the pixel difference in the sub-block is equal to the same constant vector diff_R / G / B. For example, the RGB values of all 64 pixels in this sub-block are 010 or 001, that is, they are all equal and not all are 0; Step 513: When the pixel difference of all pixels in a sub-block is the same as the pixel difference of an adjacent block, it is recorded as "neighborhood consistency" and marked as 2; the adjacent block refers to the left, top, or upper left of the sub-block, with three directions: left, top, or upper left. That is, if any one of the directions satisfies the condition, then we only need to choose one of them; Step 514: If none of the above conditions are met, record it as "complex", mark it as 3, and proceed to step 52; Step 52: Perform a second sub-mesh judgment on complex sub-blocks: Step 521: Divide the complex sub-block into a 2*2 sub-mesh to obtain several sub-blocks; A 2x2 subgrid is created by dividing an 8x8 subgrid into 16 sub-blocks using a sliding window with a step size of 2. Step 522: When the pixel difference value of all pixels in a certain sub-block is 0, record the "copy" flag as 0; Step 523: When the pixel difference value of all pixels in a sub-block is the same, record it as "monochrome" and mark it as 1; Step 524: When the pixel difference of all pixels in a sub-block is the same as the pixel difference of an adjacent block, it is recorded as "neighborhood consistency" and marked as 2; the adjacent block refers to the left, top, or top-left side of the sub-block, with three directions: left, top, or top-left. That is, steps 522 to 524 are similar in principle to steps 511 to 513; Step 525: When there are two kinds of pixel difference values for all pixels in a certain sub-block, it is recorded as "double constant" and marked as 3: that is, two kinds of constant value distribution, there are two kinds of values in the sub-block, which can be classified into the following sub-categories according to their position combination; Step 5251, "Same in the upper row"; that is, top left = top right, and bottom left = bottom right, such as Figure 5 As shown in the first picture; Step 5252, "The left column is the same"; that is, the top left = the bottom left, and the top right = the bottom right, such as... Figure 6 As shown in the second picture; Step 5253, "Diagonal same"; that is, top left = bottom right, and top right = bottom left, such as Figure 6 As shown in the third picture; Step 5254, "Only one corner is different"; that is, only the bottom right, bottom left, top left, or top right corners are different from the other triangles. Any of these conditions must be met, such as... Figure 6As shown in the fourth picture; Step 526: When there are three types of pixel differences among all pixels in a sub-block, they are recorded as "three constants" and marked as 4; that is, a three-value distribution. The sub-block has three constant values. It is classified according to which two are the same (there are a total of 6 position combinations). like Figure 6 As shown, any two of the four pixels are blue, while the other two are different colors; Step 527: When there are four types of pixel differences among all pixels in a certain sub-block, they are recorded as "four constants" and marked as 5, meaning that the four positions are different. like Figure 7 As shown, the four pixels in the four sub-blocks are all different; Step five yields the 8x8 and 2x2 containers of frame C, meaning that container C contains records and markers corresponding to two types of sub-grids.
[0021] Step 6: Divide frame B into first and second subgrids, and record and mark them; Frame B and Frame C have the same size subgrids; Step 61: Perform the first subgrid judgment on frame B; Step 611: Record as "monochrome" and mark as 1; that is, there is one and only one pixel value in the block; Step 612: Record as "Neighborhood Consistency", mark as 2, refer to step 513; Step 613: If none of the above conditions are met, record it as "complex" and mark it as 3; refer to step 514; and execute step 62; Step 62: Perform a second sub-mesh judgment on the complex sub-blocks; Step 621: Divide the coarse-grained mesh into a fine-grained mesh of size 2*2, and set the step size to 2; Step 622: Record as "monochrome" and mark as 1; that is, there is one and only one pixel value in the block; Step 623: Record as "Neighborhood Consistency" and mark as 2; refer to step 523; Step 624: Record as "double constant" and mark as 3; refer to step 524; Step 625: Record as "three constants", marked as 4; refer to step 524; Step 626: Record as "Four Constants", marked as 5; refer to step 526; Step 7: Set priorities based on the records in frame C and frame B, and extract the new frame D block by block according to the priorities; The order in which the first subgrid has the highest priority is: Priority 1: Records in the first subgrid of frame C are duplicates and marked as 0; Second priority: Records in the first subgrid of frame C as neighborhood consistent are marked as 1; Priority 3: Records in the first subgrid of frame B as neighborhood consistent are marked as 1; Priority 4: The first subgrid of frame C is recorded in monochrome and marked as 2; Priority 5: The first subgrid of frame B is recorded in monochrome and marked as 2; Priority 6: The record in the first subgrid of frame B and frame C is complex and is marked as 3; The first grid cell records complex cases as having the next lowest priority. Priority 7: Records in the second subgrid of frame C are duplicates and marked as 0; Priority 8: Records in the second subgrid of frame C as neighborhood consistent are marked as 1; Priority 9: Recorded in monochrome in the second subgrid of frame C, marked as 2; Priority 10: Records in the second word grid of frame B are neighborhood consistent and are marked as 1; Priority 11: The second subgrid of frame B is recorded in monochrome and marked as 2; Priority 12: Recorded as a double constant in the second subgrid of frame C, marked as 3; Priority 13: Recorded as a double constant in the second subgrid of frame B, marked as 3; Priority 14: Recorded as three constants in the second subgrid of frame C, marked as 4; Priority 15: Recorded as three constants in the second subgrid of frame B, marked as 4; Records with four constants in the second subgrid have low priority. Priority 16: This means that the second sub-grid is recorded as four constants, marked as 5, and single-pixel operations are performed. The above priority order can be customized; like Figure 8 In the upper half of the image, the corresponding records are extracted for each sub-block. Assuming the yellow area in the image is the first 8*8 block, this block satisfies the first priority record, which is the all-zero area, meaning the 8*8 block at the corresponding position in frame C is used as the first sub-block of frame D; the orange area represents the second priority record, and the 8*8 block at the corresponding position in frame C is used as the second sub-block of frame D; the light blue area represents the third priority, and so on. If no matching state is found, extraction continues from the second sub-grid, such as... Figure 8The lower half of the image, for example, an 8x8 sub-region, is divided into several 2x2 blocks, and then records are extracted from them. Assume the first orange area represents a quadrant with all zeros, satisfying priority 7; the first light blue quadrant represents a quadrant that is consistent with its neighbors in frame C, satisfying priority 8; the second orange area represents a quadrant with exactly one color satisfying priority 10; the second light blue area represents a quadrant with exactly one color satisfying priority 11 in frame B; the third orange area represents a quadrant in frame C... Figure 5 The seven double-length patterns (1), (2), (3), and (4) satisfy the 12th priority. The third one, light blue, indicates that it is in a B-frame. Figure 5 The seven double-length patterns (1), (2), (3), and (4) satisfy the 13th priority. The fourth orange line indicates a step in the C frame. Figure 6 The six three-constant patterns in the image satisfy priority 14. The fourth light blue pattern indicates that the six three-constant patterns in step 6 are present in frame B, satisfying priority 15; the fifth orange pattern indicates four constant patterns, which do not match in frames C and B, satisfying priority 16. A new frame D is obtained through step seven. The container of the new frame D contains the records of the first sub-mesh and the second sub-network. Step 8: Compress the container of the new frame D using entropy coding; if no inter-frame mode is hit, use the existing intra-frame compression. Intra-frame compression methods include a patented adaptive redundancy elimination image compression coding method.
[0022] And perform inter-frame entropy coding on the container of the new frame D; any existing lossless compression entropy coding method can be used, such as Huffman coding and arithmetic coding.
[0023] The efficiency of this invention is compared with that of existing conventional inter-frame compression methods, and the compression ratios when using different priorities are compared as shown in Table 1 and Table 2. Table 1. Comparison of compression ratios of the present invention and conventional inter-frame compression methods for different scenarios.
[0024] Table 2 Comparison of compression rates for different priority orders of the method of the present invention
[0025] The first priority is from priority 1 to priority 16 in step seven, and the second priority is 2, 3, 1, 4, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14, 15, 16; This invention employs a block-based strategy that can adapt to video content with different resolutions and motion complexities. It can achieve higher compression efficiency for static and weak motion areas, while maintaining stable prediction performance for complex motion areas, resulting in stronger overall robustness.
[0026] This invention utilizes both inter-frame temporal correlation and intra-frame spatial correlation. By comparing inter-frame differences at the block level, it effectively eliminates temporal redundancy between consecutive frames. It further mines spatial redundancy within a single frame by comparing intra-frame neighborhoods. Combined with the splitting and marking of pixel formats within blocks, it performs differentiated processing on different types of pixel components, enabling more thorough compression of redundant information. The overall compression ratio is significantly better than conventional lossless methods that rely solely on a single redundancy elimination method.
[0027] Based on the above-described preferred embodiments of the present invention, and through the foregoing description, those skilled in the art can make various changes and modifications without departing from the inventive concept. The technical scope of this invention is not limited to the contents of the specification, but must be determined according to the scope of the claims.
Claims
1. A lossless compression method based on block-based elimination of inter-frame redundancy, characterized in that, Includes the following steps: Step 1: Acquire the video of a specific shot and segment the video into frames; Step 2: Name adjacent frames and divide them into sub-grids; Step 3: Divide candidate blocks using the absolute error of the pixel difference between corresponding sub-blocks in the current frame and the previous frame; Step 4: Construct candidate frames using candidate blocks; Step 5: Subtract the current frame from the candidate frame to obtain the difference frame; Divide the difference frame into coarse and fine-grained sub-grids; output the recording status of sub-blocks in the sub-grid; Step 6: Divide the current frame into coarse and fine-grained sub-mesh, and output the recording status of the sub-blocks in the sub-mesh; Step 7: Set different priorities for the recording status of the current frame and the difference frame, extract the recording status from the sub-blocks of the current frame and the difference frame in order of priority, and output the new frame.
2. The lossless compression method based on block-based elimination of inter-frame redundancy according to claim 1, characterized in that, Step three specifically includes: Divide the current frame and the previous frame into several sub-blocks according to the sub-grid size; For a sub-block in the current frame, the SAD value is calculated block by block by sliding window after extending the sub-block at the same position in the previous frame by a preset number of pixels. If the SAD values are the same, they are set as the same candidate block; otherwise, the block with the smallest SAD value is selected as the similar candidate block. Output the motion vector MV and state vector of the same or similar candidate blocks.
3. The lossless compression method based on block-based elimination of inter-frame redundancy according to claim 1, characterized in that, Step five specifically includes: Step 51: Perform coarse-grained grid difference frame division; record state division based on pixel difference of sub-blocks; Step 511: When the pixel difference of all pixels in a sub-block is 0, record it as a copy state; Step 512: When the pixel difference value of all pixels in a sub-block is the same, record it as a monochrome state; Step 513: When the pixel difference of all pixels in a sub-block is the same as the pixel difference of a neighboring block, it is recorded as a neighborhood consistency. If any of steps 511 to 513 is not satisfied, perform the fine-grained mesh difference frame division in step 52.
4. The lossless compression method based on block-based elimination of inter-frame redundancy according to claim 3, characterized in that, Step 52 specifically includes: Step 521: Divide the coarse-grained mesh into a fine-grained mesh of size 2*2, and set the step size to 2; Step 522: When the pixel difference value of all pixels in a sub-block is 0, record it as a copy state; Step 523: When the pixel difference value of all pixels in a sub-block is the same, record it as a monochrome state; Step 524: When the pixel difference of all pixels in a sub-block is the same as the pixel difference of a neighboring block, it is recorded as a neighborhood consistency state. Step 525: When there are two possible pixel differences among all pixels in a sub-block, record it as a double constant state; Step 526: When there are three types of pixel differences among all pixels in a certain sub-block, it is recorded as a three-constant state; Step 527: When there are four types of pixel differences among all pixels in a certain sub-block, it is recorded as a four-constant state.
5. The lossless compression method based on block-based elimination of inter-frame redundancy according to claim 1, characterized in that, Step six specifically includes: Step 611: When the pixel difference value of all pixels in a sub-block is the same, record it as a monochrome state; Step 612: When the pixel difference of all pixels in a sub-block is the same as the pixel difference of a neighboring block, it is recorded as a neighborhood consistency state. If either step 611 or step 612 is not satisfied, perform the fine mesh difference frame division in step 62.
6. The lossless compression method based on block-based inter-frame redundancy elimination according to claim 5, characterized in that, Step 62 specifically includes: Step 621: Divide the coarse-grained mesh into a fine-grained mesh of size 2*2, and set the step size to 2; Step 622: When the pixel difference value of all pixels in a sub-block is the same, record it as a monochrome state; Step 623: When the pixel difference of all pixels in a sub-block is the same as the pixel difference of a neighboring block, it is recorded as neighborhood consistency. Step 624: When there are two possible pixel differences among all pixels in a sub-block, record it as a double constant state; Step 625: When there are three types of pixel differences among all pixels in a certain sub-block, record it as a three-constant state; Step 626: When there are four types of pixel differences among all pixels in a certain sub-block, it is recorded as a four-constant state.
7. The lossless compression method based on block-based elimination of inter-frame redundancy according to claim 1, characterized in that, The priority order is customized based on the different states of the records.
8. The lossless compression method based on block-based elimination of inter-frame redundancy according to claim 1, characterized in that, Output the container corresponding to the new frame and compress the container using entropy coding.
9. A lossless compression system based on block-based elimination of inter-frame redundancy, characterized in that, include: Memory is used to store instructions that can be executed by the processor; A processor for executing instructions to implement the lossless compression method based on block-based elimination of inter-frame redundancy as described in any one of claims 1-8.
10. A computer-readable medium storing computer program code, characterized in that, The computer program code, when executed by a processor, implements the lossless compression method based on block-based elimination of inter-frame redundancy as described in any one of claims 1-8.