CODING AND DECODING METHOD, DEVICE, ENCODER, AND DECODER FOR MULTI-VIEW VIDEO

摘要

The embodiments of the present invention provide a coding method, a decoding method for multi-view video and corresponding device, encoder, and decoder, wherein the coding method for the multi-view video comprises: minimising errors between the current coding viewpoint image and deformed viewpoint image of a front viewpoint image to obtain an optimal deformation shift; using the optimal deformation shift, camera parameters of the viewpoint and the information about the depth image of the front viewpoint image to calculate disparity information between the front viewpoint image and the current coding viewpoint image; using the disparity information and the front viewpoint image to calculate a deformed viewpoint image of the front viewpoint image, and predicting the current viewpoint image using the deformed viewpoint image as a predicting signal. The embodiments of the present invention obtain the optimal deformed shift by minimising the errors between the current coding viewpoint image and the deformed viewpoint image of the front viewpoint image, and use the optimal deformed shift to correct the disparity information, thereby increasing the precision of the deformed viewpoint image and in turn improving the image quality of multi-view video.