50results about How to "Reduce the theoretical limit entropy" patented technology

The invention relates to a prediction method used for bandwidth compression. The method comprises the following steps: acquiring first prediction residual of the current MB according to an adaptive template; acquiring a texture direction gradient value of a plurality of components of the current MB, acquiring a second prediction residual of the current MB; respectively calculating a sum of residual absolute values of the first prediction residual and a sum of residual absolute values of the second prediction residual; selecting the prediction residual of which the sum of the residual absolutevalues is relatively smaller from the first prediction residual and the second prediction residual; and transmitting the final prediction residual of the current MB and mark information of a prediction method corresponding to the final prediction residual of the current MB into a code stream. Compared with the existing method, by a prediction selection algorithm, an optimal prediction method is selected from various prediction methods, when the optimal prediction method is used for processing a complex texture image, the optimal prediction method has excellent prediction effect and high processing efficiency, and can reduce theoretical limit entropy and increase bandwidth compression rate.

The invention relates to a video coding method and a device thereof. The method comprises the following steps of: respectively coding a to-be-coded macro block according to at least two sets of codingmethods to obtain at least two coded streams; performing operation according to the at least two coded streams to form coding information corresponding to each coded stream; and obtaining a final coding method of the to-be-coded macro block according to the coding information respectively corresponding to the at least two coded streams. According to the video coding method, a code macro block iscoded by a variety of coding methods, the image code compressing rate can be improved for the image coding blocks in different scenes, and the theoretical limit entropy of the compression can be further reduced than the existing methods.

The present invention relates to an adaptive template prediction method for video compression, comprising: step 1, creating a template list; step 2, filling MB reconstruction values ​​in the template list; step 3, selecting reference pixels in the template list ; Step 4, calculating the prediction residual of the current MB according to the pixel value of the reference pixel. The invention calculates the predicted pixel value of the current predicted macroblock by creating an adaptive template, and compared with the existing method, it can improve the accuracy of calculating the predicted residual value for the complex texture area, further reduce the theoretical limit entropy, and improve the video compression rate.

The invention discloses a bandwidth compression prediction method based on a quad-tree. The prediction method comprises the following steps of: step 1, dividing a macroblock to be predicted accordingto a quad-tree algorithm; step 2, acquiring a first bit number and a first prediction residual according to the macroblock to be predicted; step 3, acquiring a second bit number and a second prediction residual according to a first sub-macroblock, a second sub-macroblock, a third sub-macroblock, and a fourth sub-macroblock; step 4, determining whether to divide the macroblock to be predicted according to the first bit number, the first prediction residual, the second bit number, and the second prediction residual; and step 5, outputting the prediction residual and the minimum value of the pixel component of each divided sub-macroblock of the macroblock to be predicted. The bandwidth compression prediction method based on the quad-tree has good prediction effect and high processing efficiency, and can reduce the theoretical limit entropy when the method is used for complex texture image processing.

The invention relates to a method for adaptive texture gradient prediction in bandwidth compression, which includes selecting N sampling methods to sample pixel components in the current MB; selecting M prediction methods to predict the current MB; and calculating the current MB respectively. The prediction residual and SAD; select the sampling mode and prediction mode of the current MB according to the SAD. The present invention calculates the prediction residual and SAD of the current prediction macroblock by defining the sampling mode of the MB. Compared with the existing methods, when the texture of the image to be compressed is more complex, for the MB at the texture boundary of the current image, according to the texture gradient principle, it does not depend on the surrounding MB of the current MB, but through the current MB The prediction residual obtained by its own texture characteristics can improve the accuracy of the prediction residual value for complex texture areas, further reduce the theoretical limit entropy, and increase the bandwidth compression rate.

The invention relates to a dual-mode selection prediction method for a complex texture in bandwidth compression, which comprises the steps of: partitioning a to-be-encoded video image into a pluralityof macro blocks, and determining a to-be-encoded pixel component; by adopting an adaptive template prediction method, selecting a first reference pixel of each currently encoded pixel in a currentlyencoded macro block in an adaptive template, and calculating to obtain a group of first predicted residuals; by adopting an adaptive rectangular window prediction method, selecting a second referencepixel of each currently encoded pixel in the currently encoded macro block in a rectangular prediction search window, and calculating to obtain one group of second predicted residuals; calculating a first subjective difference according to one group of first predicted residuals, and calculating a second subjective difference according to one group of second predicted residuals; and comparing the first subjective difference with the second subjective difference, and determining an optimal prediction method of the currently encoded macro block to obtain one group of optimal predicted residuals.The dual-mode selection prediction method uses the macro blocks as prediction units, and according to different texture characteristics in different regions of the image, the optimal prediction methods are adaptively selected, so that the dual-mode selection prediction method is better in prediction effect.

The present invention relates to a method for predicting adaptive texture gradients in bandwidth compression, comprising: selecting N sampling methods to sample the current MB, wherein the value of N is a natural number greater than 1; predicting the current MB, and obtaining the The prediction residual of the current MB; calculating the sum of the absolute value of the residual of the current MB; and determining the sampling mode of the current MB according to the sum of the absolute value of the residual. The present invention adopts multiple sampling methods for the current MB to obtain the prediction residual and SAD of the current MB. When processing a compressed image with a complex texture, for the current MB at the texture boundary of the current image to be compressed, due to the differences between the current MB and the surrounding MB Not in the same texture area, resulting in a poor correlation between the current MB and surrounding MBs, but the present invention does not depend on the MBs around the current MB based on the texture gradient principle, but obtains the prediction residual through the texture characteristics of the current MB itself, It can improve the accuracy of prediction residual value for complex texture area.

The invention relates to a multi-mode selecting and predicting method for complex textures in bandwidth compression, comprising the following steps of: dividing a to-be-coded video image into a plurality of macro blocks; determining a to-be-coded pixel component; calculating the macro block gradient of the current coding macro block; determining the gradient level of the current coding macro blockaccording to the macro block gradient of the current coding macro block in a gradient-prediction mode lookup table; determining the optimal prediction mode of the current coding macro block accordingto the gradient level; predicting the current coding macro block by using the optimal prediction mode; and calculating a set of optimal prediction residual errors of the current coding macro block. According to the video coding method as compared with the existing methods, a code macro block is coded by a variety of coding methods, the image code compressing rate can be improved for the image coding blocks in different scenes, and the theoretical limit entropy of the compression can be further reduced.

The invention relates to a prediction method based on video compression, comprising: obtaining the first residual absolute value sum of the to-be-predicted MB; obtaining the second residual absolute value sum of the to-be-predicted MB; obtaining the first standard deviation and the second standard deviation of the to-be-predicted MB; obtaining a first residual subjective sum according to the firstresidual absolute value and the first standard deviation; obtaining a second residual subjective sum according to the second residual absolute value and the second standard deviation; selecting a minimum value of the first residual subjective sum and the second residual subjective sum; and using the prediction method corresponding to the minimum value to predict the to-be-predicted MB. Accordingto the prediction method based on video compression, a variety of prediction methods are selected for predicting the to-be-compressed image, so that an optimal prediction method can be selected from avariety of prediction methods when the texture of the to-be-compressed image is complex, and therefore, the portion of the to-be-compressed image where the texture is complex can be predicted, and the compression quality of the to-be-compressed image can be guaranteed.

The invention relates to a prediction method based on video compression, which comprises the steps of obtaining first standard deviation and second standard deviation of an MB to be predicted; obtaining the first sum of absolute difference of the MB to be predicted by using first prediction residuals of the MB to be predicted; obtaining the second sum of absolute difference of the MB to be predicted by using second prediction residuals of the MB to be predicted; obtaining the first subjective difference according to the first standard deviation and the first sum of absolute difference; obtaining the second subjective difference according to the second standard deviation and the second sum of absolute difference; and determining a final prediction method for the MB to be predicted accordingto the first subjective difference and the second subjective difference. Various prediction methods are adopted for the same image to be compressed, and the image to be compressed adaptively selectsone prediction method for prediction according to the texture correlation, thereby enabling the image to be compressed to select a prediction method for prediction in a targeted manner, and greatly improving the quality of video compression.

The present invention relates to a multi-direction prediction method for block skip scanning in bandwidth compression, comprising: dividing an image into multiple MBs of the same size; and marking the multiple MBs with P marking symbols, wherein P is greater than 1 A natural number; sequentially predict the MBs corresponding to the P marker symbols in a set order to obtain a prediction residual. The present invention determines the predicted pixel value of the current pixel through the method of skipping and scanning. Compared with the existing method, more reference directions can be obtained for MB, and the theoretical limit entropy of prediction can be further reduced.

The invention relates to an adaptive template prediction method for bandwidth compression, which comprises the steps of 1, defining the epitope quantity and epitope serial numbers of an adaptive template; 2, updating the adaptive template list according to MB reconstruction values of adjacent reference directions of the current MB (Macro block); 3, carrying out adaptive texture prediction on the current MB and the MB reconstruction values in the adaptive template list to obtain prediction residuals of the current MB; and 4, judging whether all MB prediction residuals are obtained or not, if so, ending the prediction, otherwise, jumping to the step 2. The predicted pixel value of the current micro block is calculated through defining the adaptive template. Compared with the existing method,the adaptive template prediction method can improve the accuracy of solving prediction residuals for a complex texture region, thereby further reducing the theoretical limit entropy, and improving the bandwidth compression rate.