Sparse optimization of non separable vector lifting scheme for stereo image coding.

Highlights • Design of adaptive non separable vector lifting scheme for stereo image coding. • Sparse optimization of the lifting operators used with the left and right images. • Development of independent full optimization and joint optimization algorithms. • Benefits of sparse optimization techniques in the context of stereo image coding. Abstract One of the potential 3D imaging techniques relies on the use of stereoscopic systems. The great interest in these systems has resulted in huge amount of data which needs to be compressed for storage and transmission purposes. In this context, vector lifting scheme has been found to be an efficient approach for stereo image coding. For instance, the coding performance depends on the design of the involved lifting operators referred to as prediction and update filters. For this reason, while a non separable vector lifting structure is retained, we investigate different techniques for optimizing sparse criteria to design the filters used with both views. More precisely, an independent full optimization algorithm as well as a joint algorithm will be developed and studied. Simulations performed on different stereo images demonstrate the effectiveness of the proposed sparse optimization algorithms in terms of quality of reconstruction and bitrate saving. [ABSTRACT FROM AUTHOR]