FRUC Template Matching Candidate Selection for Video Decoding
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
The Frame Rate Up-Conversion (FRUC) template matching method in video coding experiences high coding complexity, particularly at the encoder due to motion vector refinement and Rate-Distortion (RD) calculations, which hinders efficient bit-rate reduction.
Innovation Solution
The method involves selecting a fixed or adaptively calculated number of template matching (TM) candidates for a temporal layer or slice during encoding or decoding, with the number of TM candidates signaled in the slice header, sequence parameter set, or picture parameter set, to reduce complexity and improve efficiency.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Reliability
If the number of template matching candidates is increased to improve prediction accuracy, then the bit-rate reduction efficiency is improved, but the encoding and decoding complexity increases
Solution Approach 1:
The patent applies dynamics by making the number of TM candidates adjustable rather than fixed. The system can dynamically select the number of candidates based on video content characteristics, temporal layer, and slice information, allowing optimization between accuracy and complexity for different coding scenarios
Solution Approach 2:
The patent changes the parameter of TM candidate number from a static value to a variable that can be adapted based on multiple factors including video data characteristics, temporal layer, and slice header information. This parameter adaptation enables the system to achieve optimal balance between prediction accuracy and computational complexity
2Productivity
If the number of template matching candidates is reduced to lower encoding and decoding complexity, then the processing time is reduced, but the bit-rate reduction efficiency deteriorates
Solution Approach 1:
The system dynamically adjusts the number of TM candidates based on actual coding conditions rather than using a fixed low value. By selecting an appropriate number of candidates adaptively, the system maintains high processing speed while preserving bit-rate reduction efficiency through intelligent parameter selection
Solution Approach 2:
The patent performs preliminary actions by pre-defining the number of TM candidates for different temporal layers and slices before actual encoding/decoding. This preliminary configuration allows the system to avoid complex real-time calculations while still achieving optimal efficiency through pre-planned candidate selection strategies
3Reliability
If the number of template matching candidates is adaptively calculated during encoding/decoding, then the bit-rate reduction efficiency is optimized, but the computational complexity increases
Solution Approach 1:
The patent applies preliminary action by calculating and configuring the number of TM candidates before the actual encoding or decoding process. The system determines appropriate candidate numbers based on video characteristics, temporal layer, and slice information in advance, avoiding complex real-time computations while still achieving optimized bit-rate reduction efficiency
Solution Approach 2:
The patent segments the video data into temporal layers and slices, and applies different TM candidate numbers to different segments based on their specific characteristics. This segmentation allows the system to optimize bit-rate reduction efficiency for each segment while controlling overall computational complexity by not applying high candidate numbers uniformly across the entire video
Data Source
AI summary
A device for video decoding may include a memory configured to store video data and a processor configured receive a bitstream including encoded video data. The processor may be configured to select a number of template matching (TM) candidates for a temporal layer or slice during the video decoding. The number of TM candidates selected are fixed prior to the video decoding, or adaptively calculated during the video decoding. The processor may be configured to generate a prediction block and residual block, based on a template matching candidate, to reconstruct the video data.