Image Encoding Device Using Adaptive Directional Prediction
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Conventional image encoding systems, such as H.264/AVC, face challenges in achieving accurate prediction and high compression rates due to directional dependence in intra-coded and inter-coded predictive encoding, particularly when dealing with images where pixels have similar signal levels in certain directions, leading to inefficiencies in compression and increased code quantities.
Innovation Solution
An image encoding device that includes an image conversion module to convert image directions, encoding modules for encoding images in various directions, and a mode selector module to compare code quantities and select the most efficient encoding method, allowing for accurate intra-coded and inter-coded prediction without altering the scanning direction or rotating the image, and utilizing interpolation formulas to include pixels inside the block for prediction.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Productivity
If conventional intra-coded predictive encoding is used with raster scanning, then compression rate is improved for images with horizontal/vertical patterns, but prediction accuracy deteriorates for images with diagonal or opposite-direction patterns
Solution Approach 1:
The patent applies dynamics by making the prediction direction adaptive rather than fixed. The encoder determines the optimal prediction direction (horizontal, vertical, diagonal, or anti-diagonal) based on the actual image content characteristics, allowing the system to dynamically adjust to different image patterns and achieve high compression rates regardless of the dominant pattern orientation.
Solution Approach 2:
The patent changes the parameter of prediction direction from a fixed raster scanning convention to a variable parameter that can take multiple values (horizontal, vertical, diagonal, anti-diagonal). This parameter change enables the system to optimize prediction accuracy for different image patterns while maintaining high compression rates across diverse image types.
2Productivity
If conventional inter-coded predictive encoding is used with fixed scanning direction, then encoding efficiency is improved for certain motion patterns, but prediction accuracy deteriorates when motion direction does not match scanning direction
Solution Approach 1:
The patent makes the inter-coded prediction direction dynamic by allowing the scanner to adapt its scanning direction based on the motion vector characteristics of the image data. The system determines the optimal scanning direction (horizontal, vertical, diagonal, or anti-diagonal) according to the actual motion patterns, enabling accurate prediction regardless of whether motion aligns with conventional raster scanning directions.
Solution Approach 2:
The patent changes the scanning direction parameter from a fixed convention to a variable parameter that adapts to motion patterns. This allows the inter-coded prediction to achieve high accuracy for various motion directions while maintaining encoding efficiency, resolving the contradiction between fixed scanning convenience and adaptive prediction accuracy.
3Device complexity
If boundary pixels only are used for prediction, then device complexity is reduced, but prediction accuracy deteriorates for pixels far from block boundaries
Solution Approach 1:
The patent extends prediction from the traditional one-dimensional boundary pixel approach to a two-dimensional approach by incorporating both boundary pixels and internal pixels of adjoining blocks into the prediction process. This dimensional expansion allows internal pixels to contribute to prediction accuracy while maintaining manageable complexity through structured processing of different pixel regions.
Solution Approach 2:
The patent applies local quality by treating boundary pixels and internal pixels differently in the prediction process. Boundary pixels provide structural constraints while internal pixels provide contextual information about local patterns. This differentiated approach optimizes prediction accuracy for pixels at different locations within a block without uniformly increasing complexity across the entire prediction mechanism.
Data Source
AI summary
An image encoding device and corresponding decoding device are disclosed for achieving a high rate of compression. The encoding device includes an image conversion module that converts the direction of the images using right-left symmetry mode convention or up-down symmetry mode conversion. Encoding modules are used to encode both the unconverted and converted images. A mode selector module compares the quantity of codes in the encoded unconverted images and the quantity of codes in the encoded converted images, and selects codes that are smaller in the quantity of codes for output.