Adaptive Filter Processing for Image Quality
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Conventional image processing methods apply uniform filter coefficients across images, failing to effectively address local noise variations, leading to image quality deterioration during encoding and decoding, especially with different orthogonal transform sizes like 4×4 and 8×8 blocks.
Innovation Solution
An image processing device that classifies images by orthogonal transform sizes and applies filter processing using coefficients tailored to the local nature of each partial image, using a Wiener Filter for noise removal, to improve image quality.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Device complexity
If uniform filter coefficients are applied across the entire image, then the device complexity is reduced and processing is simplified, but the image quality deteriorates due to inability to address local noise variations
Solution Approach 1:
The image is divided into multiple blocks based on orthogonal transform size (e.g., 4×4, 8×8 blocks). Each block is processed independently with its own filter coefficients, allowing localized noise removal while maintaining overall image quality. This segmentation enables the system to handle different noise characteristics in different regions without requiring complex global processing.
Solution Approach 2:
Different filter coefficients are applied to different image blocks according to their local characteristics. The filter means selects appropriate coefficients based on the orthogonal transform size and local noise patterns, enabling adaptive noise removal that preserves image quality while maintaining reasonable processing complexity.
2Manufacturing precision
If different filter coefficients are applied for different orthogonal transform sizes, then the image quality is improved by addressing local noise variations, but the device complexity increases
Solution Approach 1:
The filter means dynamically selects filter coefficients based on the orthogonal transform size of each block. This dynamic adaptation allows the system to optimize noise removal for each block type (4×4, 8×8, etc.) without requiring a completely separate processing path for each transform size, thus managing complexity while improving quality.
Solution Approach 2:
The filter processing parameters (coefficients) are changed according to the orthogonal transform size. By adjusting the filter coefficients to match the specific transform size and local image characteristics, the system achieves better noise removal effectiveness without requiring fundamentally different processing architectures.
3Manufacturing precision
If filter processing is applied to remove noise, then the image quality is improved, but encoding efficiency decreases due to increased processing time
Solution Approach 1:
The filter processing is applied in segmented blocks rather than to the entire image at once. This block-based approach allows parallel processing and reduces the computational burden per processing unit, improving encoding efficiency while maintaining the noise removal benefits across the entire image.
Solution Approach 2:
Filter processing is selectively applied to blocks where it provides the most benefit, rather than uniformly to all blocks. The filter means determines which blocks require noise removal based on local characteristics, applying processing only where necessary to maintain image quality while minimizing the overall processing time and preserving encoding efficiency.
Data Source
AI summary
An image decoding apparatus that includes circuitry that performs filter processing on an image using both filter coefficients and a number of filter taps being set based on an orthogonal transform size, and decodes the image obtained through the filter processing.