Multi-view Video Decoding via Depth-Based Binary Map Partitioning
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Current video codecs face challenges in efficiently encoding and decoding high-resolution or high-quality multi-view video content due to limitations in partitioning and prediction methods, leading to suboptimal data reduction and increased computational complexity.
Innovation Solution
The method involves determining a reference value from reference samples in a texture image to split a current block of a depth image into partitions, using a binary map to identify areas for prediction, and calculating prediction values based on adjacent samples and edge flags, optimizing the partitioning process for improved encoding and decoding efficiency.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Productivity
If conventional video encoding methods are used with predetermined block sizes, then the encoding process is simple, but the encoding efficiency and data reduction for multi-view video is insufficient
Solution Approach 1:
The current block is divided into multiple partitions based on depth values and binary maps, allowing different prediction methods to be applied to different regions. This segmentation enables more efficient encoding by adapting to local characteristics while maintaining manageable complexity through systematic partitioning rules.
Solution Approach 2:
The partitioning structure is dynamically determined based on the binary map and depth values rather than using fixed predetermined blocks. This dynamic adaptation allows the encoding process to optimize for each specific region, improving overall encoding efficiency for multi-view video content.
2Loss of substance
If multi-view video encoding with depth and texture pictures is implemented, then data redundancy is reduced, but computational complexity increases
Solution Approach 1:
The depth information is extracted and used to generate binary maps that guide the partitioning process. By separating the depth-based partitioning logic from the actual prediction process, the system efficiently removes redundancy while managing computation through a two-stage approach: first determining partitions via binary maps, then applying prediction methods.
Solution Approach 2:
Binary maps are generated in advance based on depth values before the actual prediction process. This preliminary action organizes the data structure and identifies partition boundaries upfront, reducing the computational burden during the prediction phase and enabling more efficient multi-view video encoding.
3Productivity
If partition areas are determined using traditional methods, then the process is straightforward, but the speed and efficiency of encoding/decoding is reduced
Solution Approach 1:
Binary maps serve as an intermediary structure that facilitates partition determination. Instead of directly computing partition boundaries from raw depth data, the system first generates binary maps that encode partition information in a compact form, making subsequent partition identification faster and more efficient while reducing the difficulty of measuring and detecting partition areas.
Data Source
AI summary
An inter-view video decoding method includes determining a reference value that is a reference for splitting a current block of a depth image corresponding to a texture image, from reference samples at a predetermined location of a reference block included in the texture image; determining a binary map of the current block by comparing the reference value and reference samples of the current block; splitting the current block into a plurality of partitions based on the binary map; and determining prediction values of the plurality of partitions based on adjacent samples of the current block and the binary map.