Abstract:3D High Efficiency Video Coding (3D-HEVC) is an extension of the HEVC standard for coding of multiview videos and depth maps. It inherits the same quadtree coding structure as HEVC for both components, which allows recursively splitting into four equal-sized coding units (CU). One of 11 different prediction modes is chosen to code a CU in inter-frames. Similar to the joint model of H.264/AVC, the mode decision process in HM (reference software of HEVC) is performed using all the possible depth levels and prediction modes to find the one with the least rate distortion cost using a Lagrange multiplier. Furthermore, both motion estimation and disparity estimation need to be performed in the encoding process of 3D-HEVC. Those tools achieve high coding efficiency, but lead to a significant computational complexity. In this article, we propose a fast mode decision algorithm for 3D-HEVC. Since multiview videos and their associated depth maps represent the same scene, at the same time instant, their prediction modes are closely linked. Furthermore, the prediction information of a CU at the depth level X is strongly related to that of its parent CU at the depth level X-1 in the quadtree coding structure of HEVC since two corresponding CUs from two neighboring depth levels share similar video characteristics. The proposed algorithm jointly exploits the inter-view coding mode correlation, the inter-component (texture-depth) correlation and the inter-level correlation in the quadtree structure of 3D-HEVC. Experimental results show that our algorithm saves 66% encoder runtime on average with only a 0.2% BD-Rate increase on coded views and 1.3% BD-Rate increase on synthesized views.

Advanced Residual Prediction In 3d-Hevc

Advanced residual prediction enhancement for 3D-HEVC

Low-complexity advanced residual prediction design in 3D-HEVC

Texture and Correlation Based Fast Intra Prediction Algorithm for HEVC

A 3D-HEVC Fast Mode Decision Algorithm for Real-Time Applications

Inter-view motion vector prediction for depth coding

Disparity Vector Based Advanced Inter-View Prediction in 3D-HEVC.

Efficient Multiple-Line-Based Intra Prediction for HEVC.

Linear adaptive search range model for uni-prediction and motion analysis for bi-prediction in HEVC

Enhanced Intra Prediction with Recurrent Neural Network in Video Coding

Deep region segmentation-based intra prediction for depth video coding

Neural Network-Based Enhancement to Inter Prediction for Video Coding

Enhanced Ctu-Level Inter Prediction with Deep Frame Rate Up-Conversion for High Efficiency Video Coding

Efficient Prediction Methods With Enhanced Spatial-Temporal Correlation for HEVC.

Deep Multi-Domain Prediction for 3D Video Coding.

Neural Network Based Inter Prediction for HEVC

Convolutional Neural Network-Based Coefficients Prediction for HEVC Intra-Predicted Residues

Interweaved Prediction for Video Coding.

Hybrid Intraprediction Based on Local and Nonlocal Correlations

Advanced Learning-Based Inter Prediction for Future Video Coding

Optimized Spatial Recurrent Network for Intra Prediction in Video Coding