Search

CN-121985114-A - Efficient verification method based on Z_delta compression algorithm

CN121985114ACN 121985114 ACN121985114 ACN 121985114ACN-121985114-A

Abstract

The invention relates to the technical field of compression algorithm verification, in particular to a high-efficiency Z-delta-based compression algorithm verification method. The method comprises the steps of restoring compressed data into uncompressed data according to recorded data compression rate, and generating test vectors according to compressed data format during test. The invention provides a verification method for realizing Z_delta compression algorithm design. The design verification personnel can efficiently generate the verification method of the test stimulus capable of specifying the compression size of the image data according to the test scene. The verification scene is greatly enriched, defect problems existing in design realization are found in advance, the SV constraint is used for automatically solving, the test vector can be generated quickly and efficiently, the size of a compression result of the test vector can be specified, the test vector has high precision, the values of certain pixels in the test vector can be specified, and the flexibility is high.

Inventors

  • YAO ZHENPING

Assignees

  • 沐曦集成电路(上海)股份有限公司

Dates

Publication Date
20260505
Application Date
20260407

Claims (9)

  1. 1. The efficient verification method based on the Z_delta compression algorithm is characterized by comprising the following steps of: S1, calculating a predicted value of each pixel; s2, obtaining a delta value of each pixel based on the predicted value of the pixel; S3, obtaining PREDICT DELTA values of each pixel based on the delta value of the pixel; S4, obtaining a symbol value according to the size of the delta value; s5, final precompressed data based on a symbol value; s6, restoring the compressed data into uncompressed data according to the recorded data compression rate; S7, generating a test vector according to the compressed data format during testing.
  2. 2. The efficient Z_delta compression algorithm-based verification method according to claim 1, wherein S7 specifically comprises the following steps: The first step, randomly compressing the data format; step two, random compression mode; thirdly, random compression rate; step four, randomly compressing the bit number of the data; Fifth, the position of the first non-FC pixel is random; sixthly, randomizing the total coding length of each pixel; Seventh, the values of each pixel stage0 are random; eighth step, the length and the numerical value of each pixel in each stage are random; Ninth, random the first non-FC pixel value; Tenth, calculating each stage data; Eleventh step, splice the data of each stage; Twelfth, the compressed data is high-order-added by 0 according to the compression rate, and 32B alignment is performed.
  3. 3. The method of claim 2, wherein the compressed data format is 0 or 1,0 represents a 16bit pattern for the pixel, and 1 represents a 24bit pattern for the pixel.
  4. 4. The method of claim 2, wherein in the second step, the compression mode is 0 or 1,0 represents the pixel is 128B mode, and 1 represents the pixel is 256B mode.
  5. 5. The efficient Z_delta compression algorithm-based verification method according to claim 2, wherein in the third step, the internal compression byte is { [0:3] } if the compression mode is 0, the internal compression byte is { [0:7] }, wherein in [0:3], 0 represents the data compression rate of 32B,1 represents 64B,2 represents 96B,3 represents 128B, and [0:7], 0 represents 32B,1 represents 64B,2 represents 96B,3 represents 128B, 4 represents 160B, 5 represents 192B, and 6 represents 224B, and 7 represents 256B.
  6. 6. The method of claim 2, wherein in the fifth step, the compressed mode is 0 and the compressed data format is 1, the location of the first non-FC pixel is { [0:31] }, otherwise the location of the first non-FC pixel is { [0:63] }.
  7. 7. The method of claim 2, wherein in the sixth step, the compression mode is 0 and the compressed data format is 0, the sum of the pixel stage values [0] to [63] plus 16 and 10 is equal to comp_bits, wherein 10 is the length of N0 in the compressed data format, and N0 is the total length of the compressed algorithm stage3 divided by 2; the compression mode is 0 and the compression data format is 1, then the sum of the pixel stage value [0] to the pixel stage value [31] plus 24 and 10 is equal to comp_bits, wherein 16 and 24 are the length of N1 in the compression data format, and N1 is the pixel of the first non_FC; the compressed mode is 1 and the compressed data format is 0, then the sum of the pixel stage value [0] to the pixel stage value [63] plus 16 and 10 equals comp_bits; the compressed mode is 1 and the compressed data format is 1, then the sum of the pixel stage value [0] to the pixel stage value [63] plus 24 and 10 equals comp_bits; The compressed data format is 0, then the pixel stage value belongs to {1, 3, 6, [8:18], 22}; the compressed data format is 1, then the pixel stage value belongs to {1, 3, [7:14], 16, [18:25], 31}.
  8. 8. The efficient Z_delta compression algorithm-based verification method according to claim 2, wherein in the seventh step, stage0 data [ I ] is traversed, if I is smaller than the position of the first non-FC pixel, stage0 data [ I ] is 0, otherwise, if I is equal to the position of the first non-FC pixel, stage0 data [ I ] is 1, otherwise, if pixel stage [ I ] is equal to 1, stage0 data [ I ] is 0.
  9. 9. The method for validating an efficient Z_delta compression algorithm as claimed in claim 3, wherein the length of each pixel occupied by each part is divided into the following parts: Only in stage0, then stage0 encodes a pixel of 0 and the first pixel of 1, where the total length of the pixel encoding is 1; Only in stage0, stage1, then stage0 is encoded as 1, and except for the first pixel encoded as 1, the total length of the pixel encoding is 3; Exists in stage0, stage1, stag, where the pixel code has a total length of 7; the case exists in the stage0+stage1+stage2+.

Description

Efficient verification method based on Z_delta compression algorithm Technical Field The invention relates to the technical field of compression algorithm verification, in particular to a high-efficiency Z-delta-based compression algorithm verification method. Background With advances in sensor technology and computer vision applications, we have been able to capture high resolution depth image data that is not only voluminous but also is filled with complex spatial information and potentially noisy, and there is a large amount of invalid or void areas in the depth image that makes it difficult for conventional compression algorithms to effectively process such data, existing compression algorithm validation methods are generally applicable to text or normal two-dimensional images, but they are not fully suited to the unique requirements of the depth image, e.g., depth images may contain a large number of nulls, and due to inaccuracy caused by sensor noise, and furthermore, real-time applications (e.g., augmented reality and autopilot) require that the compression and decompression process must be fast and resource consuming low to ensure user experience and system response speed, thus providing an efficient Z-delta compression algorithm based validation method. Disclosure of Invention The invention aims to provide a high-efficiency verification method based on a Z_delta compression algorithm, which aims to solve the problems of low efficiency and difficult verification of the traditional compression algorithm in the background technology. In order to achieve the above object, the present invention provides an efficient verification method based on z_delta compression algorithm, comprising the following steps: S1, calculating a predicted value of each pixel; s2, obtaining a delta value of each pixel based on the predicted value of the pixel; S3, obtaining PREDICT DELTA values of each pixel based on the delta value of the pixel; S4, obtaining a symbol value according to the size of the delta value; s5, final precompressed data based on a symbol value; S6, restoring the compressed data into uncompressed data according to the recorded data compression rate, and generating a test vector according to a compressed data format during testing; S7, generating a test vector according to the compressed data format during testing. As a preferable technical scheme of the invention, the step S7 specifically comprises the following steps: The first step, randomly compressing the data format; step two, random compression mode; thirdly, random compression rate; step four, randomly compressing the bit number of the data; Fifth, the position of the first non-FC pixel is random; sixthly, randomizing the total coding length of each pixel; Seventh, the values of each pixel stage0 are random; eighth step, the length and the numerical value of each pixel in each stage are random; Ninth, random the first non-FC pixel value; Tenth, calculating stage data; Eleventh step, splice the data of each stage; Twelfth, the compressed data is high-order-added by 0 according to the compression rate, and 32B alignment is performed. As a preferred technical scheme of the invention, the compressed data format in the first step is 0 or 1,0 represents a pixel in a 16bit mode, and 1 represents a pixel in a 24bit mode. As a preferable technical scheme of the invention, in the second step, the compression mode is 0 or 1,0 represents that the pixel is in 128B mode, and 1 represents that the pixel is in 256B mode. In the third step, the internal compression byte is { [0:3] }, the internal compression byte is { [0:7] } and the compression byte is 1, wherein 0 in [0:3] represents the data compression rate of 32B,1 represents 64B,2 represents 96B,3 represents 128B, 0 in [0:7] represents the data compression rate of 32B,1 represents 64B,2 represents 96B,3 represents 128B, 4 represents the data compression rate of 160B, 5 represents the data compression rate of 192B, 6 represents the data compression rate of 224B, and 7 represents the data compression rate of 256B. Relation between total length of compressed data and compression rate Compression ratioTotal length of32B0~256bit64B257~512bit96B513~768bit7128B769~1024bit160B1025~1280bit192B1281~1536bit224B1537~1792bitfail1793~2048bit In the fifth step, the compressed mode is 0 and the compressed data format is 1, the location of the first non-FC pixel is { [0:31] }, otherwise the location of the first non-FC pixel is { [0:63] }. In the sixth step, the compression mode is 0 and the compressed data format is 0, then the sum of the pixel stage value [0] to the pixel stage value [63] plus 16 and 10 equals comp_bits, wherein 10 is the bit number of N0 in the compressed data format, the value of N0 is the total length of the compression algorithm stage3 divided by 2, 16 is the bit number of N1 in the compressed data format, and the value of N1 is the value of the first non_FC pixel; The compression mode is 0 and the c