Disparity Vector Generation for Inter-View Video Prediction
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Conventional HEVC techniques face challenges in disparity vector construction, particularly when spatial neighboring blocks are unavailable within the motion estimation region, leading to inaccurate or unavailable disparity vectors, and require multiple iterations for different block units, complicating inter-view prediction.
Innovation Solution
The proposed solution simplifies and unifies disparity vector generation by using spatial disparity vectors of neighboring blocks outside the motion estimation region for inter-view prediction, refining disparity vectors at the CU level, and checking spatial neighboring blocks for accuracy, allowing parallel encoding and decoding within containing blocks.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If spatial neighboring blocks within the motion estimation region are used for disparity vector construction, then inter-view prediction accuracy is improved, but device complexity increases and multiple iterations are required for different block units
Solution Approach 1:
The patent extracts the disparity vector generation process from the complex iterative block-level operations and consolidates it at the containing block level. By taking out the disparity vector construction from individual block iterations and performing it once at the containing block level using spatial neighboring blocks, the solution reduces computational complexity while maintaining accuracy.
Solution Approach 2:
The patent makes the disparity vector generation universal by using spatial neighboring blocks outside the motion estimation region that can serve multiple block units within the containing block. This single disparity vector can be applied to multiple blocks, eliminating the need for separate iterative constructions for each block unit.
2Measurement precision
If multiple iterations are performed for different block units to construct disparity vectors, then inter-view prediction accuracy is improved, but encoding and decoding time increases
Solution Approach 1:
The patent performs preliminary action by constructing the disparity vector once at the containing block level before processing individual block units. This preliminary disparity vector construction eliminates the need for repeated iterations during block-level processing, significantly reducing encoding and decoding time while preserving prediction accuracy.
Solution Approach 2:
The patent merges multiple iterative disparity vector constructions into a single operation at the containing block level. By combining the disparity vector generation for multiple block units into one unified process using spatial neighboring blocks, the solution reduces processing time while maintaining the necessary prediction accuracy.
3Productivity
If spatial neighboring blocks outside the motion estimation region are used for disparity vector generation, then processing efficiency is improved, but the availability of accurate spatial references may be reduced
Solution Approach 1:
The patent transitions from the traditional two-dimensional motion estimation region to a three-dimensional spatial context by utilizing neighboring blocks outside the motion estimation region. This dimensional expansion provides additional spatial references that improve efficiency while the selection process ensures reliability by choosing appropriate neighboring blocks.
Data Source
AI summary
In one example, a video coder (e.g., a video encoder or a video decoder) is configured to determine that a current block of video data is coded using a disparity motion vector, wherein the current block is within a containing block, based on a determination that a neighboring block to the current block is also within the containing block, substitute a block outside the containing block and that neighbors the containing block for the neighboring block in a candidate list, select a disparity motion vector predictor from one of a plurality of blocks in the candidate list, and code the disparity motion vector based on the disparity motion vector predictor. In this manner, the techniques of this disclosure may allow blocks within the containing block to be coded in parallel.