Depth Compression for Multi-View Sequences Using 3-D Mesh Representation

3-D 메쉬 모델을 이용한 다시점 영상의 깊이 정보 압축

In this work, we propose a compression algorithm for depth images, which are obtained from multi-view sequences. The proposed algorithm represents a depth image using a 3-D regular triangular mesh and predictively encodes the mesh vertices using a linear prediction scheme. The prediction errors are encoded with a arithmetic coder. Simulation results demonstrate that the proposed algorithm provides better performances than the JPEG2000 lossless coder.