Hierarchical Temporal Analysis for Video Coding Quality
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing video coding methods face challenges in efficiently managing motion complexity, particularly when scene cuts occur, leading to high coding costs and poor visual quality due to excessive non-connected pixels and energy distribution issues in high-frequency pictures.
Innovation Solution
A method that performs hierarchical temporal analysis with motion compensated temporal filtering, dynamically switching between intra coding and temporal filtering based on correlation thresholds, and using reverse scaling to adapt low-frequency pictures, allowing for constant GOP size and improved quality distribution.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Productivity
If motion compensated temporal filtering is used for video coding, then temporal redundancy is exploited and compression efficiency is improved, but when scene cuts occur the quality distribution becomes poor and coding cost increases due to excessive non-connected pixels
Solution Approach 1:
The patent dynamically adjusts the GOP (Group of Pictures) size based on motion complexity detection. When high motion complexity is detected (indicating potential scene cuts), the system reduces GOP size to prevent quality degradation. This dynamic adaptation allows the system to maintain high compression efficiency during normal sequences while quickly responding to scene changes to preserve quality distribution.
Solution Approach 2:
The patent implements a feedback mechanism that monitors motion complexity metrics throughout the video sequence. This feedback is used to detect scene cuts and trigger GOP size adjustments. The system continuously evaluates motion vectors and complexity measures, and when thresholds are exceeded, it modifies the coding structure to maintain quality, thereby resolving the contradiction between compression efficiency and quality distribution.
2Reliability
If variable GOP size is used to adapt to motion complexity, then quality distribution is improved, but device complexity and management difficulty increase
Solution Approach 1:
The patent changes the parameter of GOP size dynamically based on detected motion complexity. Rather than using a fixed GOP size or complex variable GOP structures, the system adjusts a key parameter (GOP size) in response to measured conditions. This approach improves quality distribution while keeping the management mechanism relatively simple, as it relies on threshold-based detection and straightforward parameter adjustment.
Solution Approach 2:
The system introduces dynamics into the GOP structure by allowing its size to vary based on real-time motion complexity assessment. This dynamic adaptation enables the system to respond to scene changes without requiring complex pre-planned variable GOP structures, thereby improving quality distribution while maintaining manageable device complexity through rule-based adaptation.
3Measurement precision
If motion estimation is performed at full resolution, then motion accuracy is improved, but computational cost and processing time increase significantly
Solution Approach 1:
The patent segments the video processing into different resolution levels. Motion estimation is performed at reduced resolutions to obtain initial motion vectors, which are then refined at higher resolutions. This multi-resolution segmentation approach maintains adequate motion accuracy while significantly reducing the computational burden and processing time compared to full-resolution motion estimation throughout.
Solution Approach 2:
The system performs preliminary motion estimation at lower resolutions before finalizing motion vectors at full resolution. This preliminary action provides a good initial estimate that reduces the search space for subsequent refinement steps, thereby maintaining motion accuracy while reducing overall processing time and computational cost.
Data Source
AI summary
The method realizes a motion compensated temporal filtering (MCTF), the temporal filtering being replaced by an intra mode coding to obtain at least one low (L) or high (H) frequency picture if the current picture has a level of correlation with a lower previous picture at a threshold, the low frequency pictures obtained (L) being thus scaled to be adapted, at the energy level, to the pictures obtained by motion compensated temporal filtering, and comprises, at the end of analysis:a selection of the pictures obtained by intra coding of a picture of a low decomposition level with the additional condition, for the high frequency pictures, that this picture is derived itself from an intra coding.a calibration of the picture selected by carrying out at least one reverse step of the scaling step.The applications relate to video compression with temporal prediction.