Variable Block Video Encoding via Tree Partitioning
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing video compression technologies face inefficiencies in encoding high-resolution images due to the use of fixed macroblock sizes, which limits the size and prediction or transform units of subblocks, resulting in suboptimal encoding efficiency.
Innovation Solution
A method and apparatus that allow for encoding and decoding of videos using variable-sized blocks and subblocks, where the video is split into subblocks of various sizes based on a tree structure, enabling efficient encoding and decoding of block partition information, and allowing for intra or inter prediction modes for each subblock.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Productivity
If fixed macroblock sizes are used for encoding, then the encoding process is simple and standardized, but the encoding efficiency for high-resolution images is insufficient
Solution Approach 1:
The patent divides the video image into multiple layers of blocks, where each layer can have different block sizes. The video is first divided into macroblocks, then further subdivided into subblocks, and can be recursively divided into smaller blocks. This multi-layer segmentation allows the system to adapt block sizes to the content characteristics of different regions, improving encoding efficiency for high-resolution images while maintaining a structured approach to complexity management.
2Adaptability or versatility
If fixed macroblock sizes are used, then the prediction and transform units are limited, but the encoding process remains straightforward
Solution Approach 1:
The patent introduces dynamic block size selection where the block size for prediction and transform operations can vary depending on the content characteristics. Each block can be independently divided into subblocks of different sizes based on gradient calculations and variance analysis. This dynamic adaptation allows the system to use larger blocks for smooth regions and smaller blocks for high-frequency regions, enhancing prediction unit flexibility while using objective criteria to manage the complexity of block partitioning.
3Productivity
If variable block sizes are implemented, then encoding efficiency improves, but the complexity of managing different block sizes increases
Solution Approach 1:
The patent changes the parameter of block size from fixed to variable by introducing a multi-layer block structure. The system calculates gradient magnitudes and variances for different block sizes and selects the optimal block size for each region based on these parameters. This parameter-based approach allows the system to achieve variable block sizes for improved compression efficiency while using objective mathematical criteria to automatically determine the appropriate block size, reducing the subjective complexity of managing different block sizes.
Data Source
AI summary
A method performed by an apparatus of decoding a video in the unit of blocks, includes: extracting and decoding a first information corresponding to a minimum block size and a second information corresponding to a difference between the minimum block size and a maximum block size, from a bitstream; calculating the maximum block size based on the first information corresponding to the minimum block size and the second information corresponding to the difference; decoding, from the bitstream, partition information of a square current block having the derived maximum block size to identify one or more subblocks in the current block which have sizes equal to or larger than the minimum block size; decoding a prediction mode information of each of the subblocks, the prediction mode information indicating whether a subblock is inter-predicted or intra-predicted; and reconstructing each of the subblocks in the current block by predicting each of the subblocks.