Intra-frame fast coding algorithm for 3D-HEVC depth map based on two-dimensional entropy and variance
| Autor: | Huang Hongfei, Yao Wangze, Hua Zhang, ZhengXuan Wu |
|---|---|
| Rok vydání: | 2021 |
| Předmět: |
Computer science
020206 networking & telecommunications 02 engineering and technology Image segmentation Intra-frame Depth map Algorithmic efficiency 0202 electrical engineering electronic engineering information engineering Benchmark (computing) Entropy (information theory) 020201 artificial intelligence & image processing Algorithm Time complexity Coding (social sciences) |
| Zdroj: | 2021 IEEE 5th Advanced Information Technology, Electronic and Automation Control Conference (IAEAC). |
| Popis: | The 3D-High Efficiency Video Coding Standard (HEVC)[1] improves the coding quality by employing multi-viewpoint and depth map coding formats. However, the 3D-HEVC algorithm has high time complexity due to the determination process of the coding unit (CU) size and the recursive rate-distortion cost calculation. To address this problem, we present a fast algorithm, which aims to reduce the time of not only deep image segmentation decision but also the powerful RD optimization. Through fast decision-making on the 3D-HEVC depth map coding unit (CU), which is based on two-dimensional entropy of image and variance, the coding efficiency is improved rapidly. The benchmark experimental on the basic database show that the coding time of the proposed algorithm is reduced by 36.7%. |
| Databáze: | OpenAIRE |
| Externí odkaz: |
|