Disparity Vector Derivation Using Depth Sample Subsets
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Current 3D video coding systems, such as 3DV-HTM, face high computational complexity in deriving disparity vectors, particularly when determining the maximum disparity vector for inter-view motion prediction, which requires comparing 256 depth values within a 16x16 depth block.
Innovation Solution
The method simplifies disparity vector derivation by using a subset of depth samples, excluding the center sample, to determine a derived disparity vector based on maximum, average, median, or most frequent depth values, reducing the number of comparisons and improving efficiency.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If the maximum disparity vector is derived by comparing all depth values in a 16x16 depth block, then the accuracy of inter-view motion prediction is improved, but the computational complexity increases significantly
Solution Approach 1:
The patent divides the 16x16 depth block into four 8x8 sub-blocks and processes each sub-block separately to find local maximum depth values. This segmentation reduces the computational burden from comparing 256 values to comparing 64 values across four smaller regions, while still capturing the depth variations needed for accurate disparity vector derivation.
Solution Approach 2:
The patent extracts and uses only the maximum depth value from each 8x8 sub-block rather than processing all 256 depth values. This extraction approach identifies the critical information (maximum depth values representing nearest objects) while discarding redundant data, significantly reducing computational complexity while maintaining prediction accuracy.
2Ease of manufacture
If conventional video coding techniques are applied to each single-view video sequence independently, then the coding process is simplified, but the coding efficiency deteriorates due to ignoring inter-view redundancy
Solution Approach 1:
The patent implements a unified 3D video coding framework that handles multiple views simultaneously, enabling the system to exploit inter-view redundancy across different views. This multi-functional approach allows the same coding process to process base view and dependent views while utilizing depth information and inter-view predictions, improving overall coding efficiency without significantly complicating the implementation.
3Adaptability or versatility
If more cameras are used to generate multi-view video with a large number of video sequences, then the viewing experience is improved, but the storage space and transmission bandwidth requirements increase
Solution Approach 1:
The patent changes the representation parameters by introducing depth maps as an additional data type alongside traditional video sequences. By encoding depth information separately and using it for inter-view prediction, the system can represent multiple views more compactly, reducing the actual storage and bandwidth requirements while maintaining the ability to generate multiple viewpoints.
Solution Approach 2:
The patent uses depth maps as a compact representation that can be copied and reused across multiple views. Instead of storing complete video sequences for all views, the system stores depth maps that can be leveraged to synthesize and predict dependent views, effectively copying the essential geometric information across multiple perspectives to reduce redundant data storage.
Data Source
AI summary
A method and apparatus for three-dimensional video encoding or decoding using the disparity vector derived from an associated depth block are disclosed. The method determines an associated depth block for a current texture block and derives a derived disparity vector based on a subset of depth samples of the associated depth block. The subset contains less depth samples than the associated depth block and the subset excludes a single-sample subset corresponding to a center sample of the associated depth block. The derived disparity vector can be used as an inter-view motion (disparity) vector predictor in Inter mode, an inter-view (disparity) candidate in Merge mode or Skip mode. The derived disparity vector can also be used to locate a reference block for inter-view motion prediction in Inter mode, inter-view candidate in Merge or Skip mode, inter-view motion prediction, inter-view disparity prediction, or inter-view residual prediction.