Image Encoding Disparity Vector Prediction Using Block Inversion
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Current image encoding methods, such as MPEG4-AVC, face limitations in prediction efficiency due to low correlation between adjacent blocks, especially when the correlation is low on the boundary of the encoding target block, leading to reduced encoding efficiency.
Innovation Solution
The proposed solution involves an image encoding and decoding method that uses a disparity vector detection unit to identify a signal with high correlation to the encoding target block, generating a prediction signal by inverting a reference image in horizontal or vertical directions, and encoding the difference signal with the disparity vector and inversion direction information, thereby improving prediction efficiency.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Productivity
If intra-prediction is used with adjacent decoded image blocks, then encoding efficiency is improved through correlation between adjacent pixels, but prediction accuracy deteriorates when correlation is low on block boundaries
Solution Approach 1:
The patent creates virtual copies of the decoded image block by inverting pixel values (complement operation) to generate alternative prediction candidates. This copying approach provides additional prediction options beyond the original block, enabling selection of the most accurate prediction source even when adjacent block correlation is low.
Solution Approach 2:
The patent transforms the prediction approach by changing the parameter state of the reference block through inversion operations. By creating inverted versions of the decoded block and evaluating multiple inversion states, the system adapts to varying correlation conditions and selects the optimal prediction configuration.
2Measurement precision
If disparity vectors are transmitted to specify reference image positions, then prediction accuracy is improved by selecting high-correlation regions, but encoding complexity increases due to additional data transmission
Solution Approach 1:
The patent extracts only the essential inversion direction information from the complex disparity vector concept, eliminating the need to transmit full disparity vectors. By focusing on the inversion state rather than precise position mapping, the system maintains prediction accuracy while significantly reducing data transmission requirements.
Solution Approach 2:
Instead of transmitting disparity vectors to locate reference blocks, the patent creates inverted copies of the current decoded block and uses these copies directly for prediction. This copying mechanism eliminates the need for complex inter-block disparity calculations and data transmission.
3Measurement precision
If multiple prediction modes are evaluated to find the best match, then prediction accuracy is improved, but processing time increases
Solution Approach 1:
The patent performs a limited set of inversion operations (typically 2-4 inversion states) rather than exhaustively searching all possible reference blocks. This partial action approach evaluates sufficient prediction candidates to achieve high accuracy while avoiding excessive processing time associated with complete searches.
Data Source
AI summary
A disparity vector detection unit retrieves a signal having a high correlation with an encoding target block and acquire the signal having the high correlation and a disparity vector which is a disparity in a screen of the encoding target block by using a local decoded image of a block in the same image signal which was previously encoded, with respect to the encoding target block. A disparity prediction signal generation unit generates a prediction signal in accordance with the disparity vector. The disparity vector detection unit includes a DC calculation unit operative to predict a DC component of an image signal of the encoding target block from a neighboring decoded image, and a disparity reference DC calculation unit operative to calculate a DC component of a prediction signal represented by the disparity vector.