Convex Hull Video Encoding with Shot-Based Quantizer Prediction
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Conventional video encoding methods face high computational demands and inefficiencies, leading to increased coding time and reduced coding speed, which strain video processing infrastructure and affect network streaming performance.
Innovation Solution
A convex hull encoding method that splits video clips into shots with similar scenes, applies down-sampling and up-sampling, calculates quality metrics, selects convex hull points, and predicts quantizer parameters to optimize encoding parameters, reducing computational complexity while maintaining coding efficiency.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Manufacturing precision
If conventional encoding methods are used to improve coding efficiency, then coding efficiency is improved, but coding speed becomes slower and computational loading increases
Solution Approach 1:
The video clip is divided into multiple video shots based on scene changes, and each shot is processed independently. This segmentation allows parallel processing of different shots, improving overall coding speed while maintaining quality through shot-level optimization
Solution Approach 2:
The method performs preliminary actions by pre-calculating convex hull points and operation point series from reference video shots before actual encoding. Quantizer parameters are predicted in advance using these pre-computed data structures, eliminating the need for complex real-time calculations during encoding and significantly improving coding speed
2Manufacturing precision
If complex encoding methods are used to obtain good coding efficiency, then coding efficiency is improved, but computational loading to the computing device increases
Solution Approach 1:
Complex calculations for convex hull points, operation point series, and quantizer parameter predictions are performed in advance during a preparation phase. This shifts computational burden from the encoding stage to a separate preprocessing stage, reducing real-time computational loading while preserving coding efficiency
Solution Approach 2:
The method uses reference video shots as copies to build convex hull points and operation point series. Instead of performing complex optimizations on every frame, the system replicates and adapts parameters from reference shots, significantly reducing computational requirements while maintaining quality
Data Source
AI summary
A method of forming a convex hull encoding method is provided. The method includes the step of providing a source video clip; splitting the source video clip by a video splitter with a similar scene; slicing each video shots to a preset length video shots; providing down-sampling process, encoding process, decoding process and up-sampling process to obtain the analysis video shots; calculating a quality index between the video shots and the analysis video shots to obtain quality metrics; selecting convex hull points and calculating slopes between the convex hull points; picking operation points with similar slopes to form operation point series and predicting quantizer parameters; encoding the source video clip by the quantization parameters and the corresponding resolutions to obtain a compressed video.