High-performance depth map coding for 3D-AVC
| Autor: | Jinmi Kang, Ki-Dong Chung |
|---|---|
| Rok vydání: | 2015 |
| Předmět: |
Computational complexity theory
Computer science business.industry Depth map coding 020206 networking & telecommunications 02 engineering and technology Coding tree unit Quality enhancement Depth map Algorithmic efficiency Signal Processing 0202 electrical engineering electronic engineering information engineering 020201 artificial intelligence & image processing Computer vision Artificial intelligence Electrical and Electronic Engineering business Context-adaptive binary arithmetic coding Coding (social sciences) |
| Zdroj: | Signal, Image and Video Processing. 10:1017-1024 |
| ISSN: | 1863-1711 1863-1703 |
| DOI: | 10.1007/s11760-015-0853-6 |
| Popis: | Achieving high-performance coding for a depth map is one of the most important challenges in 3D video coding. In this paper, a quality enhancement method is proposed to accomplish better coding efficiency. A new technique separating contour and flat regions is designed, and a contour-aware quality enhancement algorithm is presented to improve depth map quality. We also propose a fast mode decision process to reduce computational complexity. The proposed fast algorithm uses similarity between texture video and depth map coding. The encoding process for a depth map is terminated early by using coded information from a texture video. Experimental results show that the quality of the depth map is improved by 0.11–0.59 dB, which translates into a bit rate saving of 2.19–8.19 %. The proposed fast algorithm saves encoding time, on average, by 36.4 %. |
| Databáze: | OpenAIRE |
| Externí odkaz: |
|
Full text from SpringerLink