Intra Prediction Mode Selection Using Frequency Domain Direction Analysis
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Conventional intra prediction methods in image coding, such as H.264/MPEG-4 AVC, require repeated operations to determine the optimal prediction mode, leading to increased complexity and computational load without significantly improving the compression rate.
Innovation Solution
The method involves frequency transforming an input block, determining the direction of pixels using coefficients, and selecting a predetermined intra prediction mode that best matches the block's features, thereby reducing the number of necessary operations by performing intra prediction only in the most suitable modes.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If conventional intra prediction methods perform repeated operations to determine optimal prediction mode, then coding accuracy is improved, but computational complexity and processing time increase
Solution Approach 1:
The patent performs frequency transformation and determines pixel directions before selecting intra prediction modes. By pre-processing the input block to extract directional information, the method avoids repeated complex operations during mode selection, thereby reducing computational complexity while maintaining coding accuracy through informed mode selection based on pre-analyzed block characteristics
Solution Approach 2:
The patent extracts directional information from frequency transformed coefficients and uses this extracted feature to guide intra prediction mode selection. By taking out only the essential directional characteristics rather than performing exhaustive mode evaluations, the method reduces computational burden while preserving coding precision
2Measurement precision
If conventional intra prediction methods test all available modes, then optimal mode selection is achieved, but processing speed decreases
Solution Approach 1:
The patent performs frequency transformation and directional analysis before mode selection, preparing the necessary information in advance. This preliminary action enables rapid mode selection based on pre-computed directional features rather than testing all modes, thereby improving processing speed without sacrificing mode selection accuracy
Solution Approach 2:
Instead of evaluating all intra prediction modes, the patent performs partial action by selecting modes based on directional information derived from frequency transformed coefficients. This selective approach evaluates only the most relevant modes for each block type, significantly improving processing speed while maintaining accurate mode selection through intelligent filtering
3Device complexity
If frequency transformation and directional analysis are performed, then intra prediction complexity is reduced, but additional computational steps are introduced
Solution Approach 1:
The patent uses frequency transformation, which is already a standard component in video coding pipelines, to serve multiple purposes: both for compression and for extracting directional information for mode selection. This multi-functional use of existing infrastructure reduces implementation complexity despite adding directional analysis, as it leverages already-present computational resources
Data Source
AI summary
A method and a device for intra prediction coding and decoding of an image in which the intra prediction according to a selected intra prediction mode based on a direction of an input image is performed are provided. In the method and the device, an input block is frequency transformed, a direction of pixels in the input block is determined, and the intra prediction is performed according to an intra prediction mode that has the most similar direction to the determined direction of the pixels.