Method for image prediction of multi-view video codec and computer readable recording medium therefor

Abstract

Provided are a method for image prediction of a multi-view video codec capable of improving coding efficiency, and a computer readable recording medium therefor. The method for image prediction of a multi-view video codec includes partitioning an image to a plurality of base blocks, acquiring information of reference images which are temporally different, acquiring information of reference images which have different views, and predicting a target block based on the acquired information. Accordingly, an image that is most similar to an image of a view to be currently compressed is generated using multiple images of different views, so that coding efficiency can be improved.

What is claimed is: 
     
       1. A method for image prediction of a multi-view video codec, which encodes multi-view images comprising a base-view target-view image and at least one different-view image which is captured from a different viewpoint from a viewpoint capturing the base-view target-view image, the base-view target-view image and the different-view image being captured at a same time, the method comprising:
 partitioning the base-view target-view image into a plurality of base blocks to be currently encoded; 
 predicting a base block among the base blocks based on the different-view image and a different-time image Which which is captured at a time different from a time when the base-view target-view image is captured, the base-view target-view image and the different-time image being captured from a same viewpoint; 
 acquiring a difference between a prediction result of the base block and the base block, and encoding the difference; and 
 generating a bitstream comprising the encoded difference, 
 wherein the predicting the base block comprises using at least one of a first residual adjusted using a second residual and the second residual adjusted using the first residual, and 
 wherein the first residual is a difference between the base-view target-view image and the different-time image, and the second residual is a difference between the base-view target-view image and the different-view image. 
 
     
     
       2. The method of  claim 1 , wherein information about the first residual and the second residual is contained in a macroblock layer or a higher layer than the macroblock layer. 
     
     
       3. The method of  claim 2 , wherein the higher layer than the macroblock layer is a slice header extension (SHE), a picture parameter set extension (PPSE), or a sequence parameter set extension (SPSE). 
     
     
       4. The method of  claim 1 , wherein the adjusted first residual is a result of adding the second residual to the first residual or subtracting the second residual from the first residual, and the adjusted second residual is a result of adding the first residual to the second residual or subtracting the first residual from the second residual. 
     
     
       5. The method of  claim 1 , wherein the predicting the base block based on the different-view image comprises performing weighted prediction by applying a weighting value to at least one of the base-view target-view image and the different-view image. 
     
     
       6. The method of  claim 5 , wherein the generating the bitstream comprises including the weighting value in the bitstream. 
     
     
       7. The method of  claim 5 , wherein in the predicting the base block, a same weighting value is applied to both the base-view target-view image and the different-view image. 
     
     
       8. The method of  claim 5 , wherein in the predicting the base block, different weighting values are applied to the base-view target-view image and the different-view image, respectively. 
     
     
       9. The method of  claim 5 , wherein the generating the bitstream comprises including information about the weighted prediction in the bitstream. 
     
     
       10. The method of  claim 1 , wherein the different-time image is an image which is the most similar to the base-view target-view image among a plurality of different-time images with respect to the base-view target-view image, and the different-view image is an image which is the most-similar to the base-view target-view image among a plurality of different-view images with respect to the base-view target-view image. 
     
     
       11. A non-transitory computer-readable recording medium storing a program for executing a method for image prediction of a multi-view video codec, which encodes multi-view images comprising a base-view target-view image and at least one different-view image which is captured from a different viewpoint from a viewpoint capturing the base-view target-view image, the base-view target-view image and the different-view image being captured at a same time, the method comprising:
 partitioning the base view target-view image into a plurality of base blocks to be currently encoded; 
 predicting a base block among the base blocks based on the different-view image and a different-time image which is captured at a time different from a time when the base view target-view image is captured, the base view target-view image and the different-time image being captured from a same viewpoint; 
 acquiring a difference between a prediction result of the base block and the base block, and encoding the difference; and 
 generating a bitstream comprising the encoded difference, 
 wherein the predicting the base block comprises using at least one of a first residual adjusted using a second residual and a second residual adjusted using the first residual, and 
 wherein the first residual is a difference between the base-view target-view image and the different-time image, and the second residual is a difference between the base-view target-view image and the different-view image. 
 
     
     
       12. A method for image prediction of a multi-view video codec, which encodes multi-view images comprising a base-view target-view image and at least one different-view image which is captured from a different viewpoint from a viewpoint capturing the base-view target-view image, the base-view target-view image and the different-view image being captured at a same time, the method comprising:
 partitioning the base-view target-view image into a plurality of base blocks to be currently encoded; 
 predicting a base block among the base blocks based on the different-view image or and a different-time image which is captured at a time different from a time when the base-view target-view image is captured, the base-view target-view image and the different-time image being captured from a same viewpoint; 
 acquiring a difference between a prediction result of the base block and the base block, and encoding the difference; and 
 generating a bitstream comprising the encoded difference, 
 wherein in case where the base block is predicted by using the different-time image and the different-view image as reference images, and the predicting the base block comprises performing weighted prediction by applying at least one weighting value to the base-view image and the different-view image reference images. 
 
     
     
       13. The method of  claim 12 , wherein the generating the bitstream comprises including the weighting value in the bitstream. 
     
     
       14. The method of  claim 12 , wherein in case where the base block is predicted by using the different-view image, a same weighting value is applied to both the base-view target-view image and the different-view image. 
     
     
       15. The method of  claim 12 , wherein in case where the base block is predicted by using the different-view image, different weighting values are applied to the base-view target-view image and the different-view image, respectively. 
     
     
       16. The method of  claim 12 , wherein in case where the base block is predicted by using the different-view image, the generating the bitstream comprises of including information about the weighted prediction in the bitstream. 
     
     
       17. The method of  claim 12 , wherein the different-time image is an image which is the most similar to the base-view target-view image among a plurality of different-time images with respect to the base-view target-view image, and the different-view image is an image which is the most-similar to the base-view target-view image among a plurality of different-view images with respect to the base-view target-view image. 
     
     
       18. A method for image decoding of a multi-view video codec, which decodes multi-view images comprising a target-view image and at least one different-view image which is captured from a different viewpoint from a view point capturing the target-view image, the target-view image and the different-view image being captured at a same time, the method comprising:
 partitioning the target-view image into a plurality of base blocks to be currently decoded;   predicting a base block among the base blocks based on the different-view image and a different-time image which is captured at a time different from a time when the target-view image is captured, the target-view image and the different-time image being captured from a same viewpoint;   decoding a difference between a prediction result of the base block and the base block; and   reconstructing the base block based on the prediction result and the difference,   wherein the base block is predicted by using the different-time image and the different-view image as reference images, and the predicting the base block comprises performing weighted prediction by applying at least one weighting value to the reference images.    
     
     
       19. The method of claim 18, wherein the predicting the base block comprises decoding the weighting value from a bitstream.  
     
     
       20. The method of claim 18, wherein the predicting the base block comprises decoding information about the weighted prediction from a bitstream.