Intra Prediction Mode Selection in Video Compression
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Current methods for selecting intra prediction modes in video compression, such as the full search method, are computationally expensive and do not achieve maximum compression performance due to reliance on fewer mode candidates and rough evaluation of matching errors.
Innovation Solution
An alternative full search method that eliminates useless modes before calculating costs, using estimated sums of predicted blocks to reduce computational complexity while maintaining peak signal-to-noise ratio and bit-rate performance.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Reliability
If the conventional full search method is used to evaluate all prediction modes, then the compression performance is maximized, but the computational complexity increases significantly
Solution Approach 1:
The patent applies preliminary action by calculating estimated sums of predicted blocks for all prediction modes before the actual cost calculation. This preliminary estimation allows the system to identify and eliminate modes with high estimation errors early in the process, avoiding the need to perform full cost calculations for all modes. The estimated sum is computed using a simplified formula that leverages boundary pixel values, providing a quick filter mechanism that preserves compression performance while reducing computational complexity.
Solution Approach 2:
The patent extracts and eliminates useless prediction modes from the full search process by comparing their estimated sums against a threshold derived from the minimum estimated sum. Modes that fail to meet the threshold criterion are extracted from the set of modes requiring full cost calculation. This extraction principle allows the system to focus computational resources only on promising modes, significantly reducing the number of operations while maintaining the ability to achieve optimal compression performance.
2Device complexity
If fewer prediction mode candidates are used to reduce complexity, then the computational complexity decreases, but the compression performance deteriorates
Solution Approach 1:
The patent changes the parameter evaluation approach by introducing estimated sum calculations as a preliminary filtering criterion. Instead of directly reducing the number of modes evaluated, the system transforms the evaluation parameter from full cost calculation to a simpler estimated sum metric. This parameter change enables the system to maintain a comprehensive mode search space while effectively reducing computational burden through efficient filtering based on the estimated sum threshold.
3Measurement precision
If the full cost calculation is performed for all prediction modes, then the mode selection accuracy is improved, but the processing time increases
Solution Approach 1:
The patent performs preliminary action by calculating estimated sums for all prediction modes before conducting full cost calculations. This preliminary step establishes a threshold based on the minimum estimated sum, allowing the system to identify modes worthy of full cost calculation. By performing this preliminary filtering action, the system maintains mode selection accuracy for promising modes while avoiding time-consuming full calculations for modes that are unlikely to be optimal.
Solution Approach 2:
The patent applies skipping by bypassing the full cost calculation step for prediction modes that fail the estimated sum threshold test. Instead of rushing through all modes with complete evaluation, the system intelligently skips the time-consuming full cost calculation for modes that are unlikely to provide optimal results. This selective skipping preserves accuracy for critical modes while dramatically reducing overall processing time.
Data Source
AI summary
A method and apparatus for selecting a coding mode of intra prediction in video compression. The method adapts an alternative full search (AFS) algorithm which reduces the computations of a conventional full search (FS) method. In the method, estimated sums of predicted blocks are obtained without actually forming predictors. By comparing the estimated sums of all modes with a low bound threshold, the undesired modes are quickly eliminated, and thus the number of predictor calculations and cost evaluations that require intensive computations are decreased.