Proceedings of the 27th annual conference on Computer graphics and interactive techniques
High-quality video view interpolation using a layered representation
ACM SIGGRAPH 2004 Papers
View synthesis prediction for multiview video coding
Image Communication
Efficient depth map compression exploiting segmented color data
ICME '11 Proceedings of the 2011 IEEE International Conference on Multimedia and Expo
Efficient Prediction Structures for Multiview Video Coding
IEEE Transactions on Circuits and Systems for Video Technology
| Hi-index | 0.00 |
Advanced 3D video systems employ multi-view video-plus-depth data to support the free-viewpoint navigation and comfortable 3D viewing; thus efficient depth map coding becomes an important issue. Unlike the color image, the depth map has a property that depth values of the inner part of an object are monotonic, but those of object boundaries change abruptly. Therefore, residual data generated by prediction errors around object boundaries consume many bits in depth map coding. Representing them with segment data can be better than the use of the conventional transformation around the boundary regions. In this paper, we propose an efficient depth map coding method using a residual segmentation instead of using transformation. The proposed residual segmentation divides residual data into two regions with a segment map and two mean values. If the encoder selects the proposed method in terms of rates, two quantized mean values and an index of the segment map are transmitted. Simulation results show significant gains of up to 10 dB compared to the state-of-the-art coders, such as JPEG2000 and H.264/AVC.