Adaptive Image Filter for Non-Uniform Feature Filtering
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Conventional image filtering techniques, such as ALF, face difficulties in performing appropriate filtering when image features are not uniform within a unit area, as they use invariant filter coefficients for each unit area.
Innovation Solution
An adaptive image filter that calculates pixel values in the output image using a filter coefficient vector, which can be changed based on the target area in the input image or the target pixel in the output image, allowing for more appropriate filtering by adjusting coefficients according to the specific features of each area.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Ease of operation
If ALF performs filtering with invariant filter coefficients for each unit area, then the filtering process is simple and fast, but it is difficult to perform appropriate filtering when image features are not uniform within a unit area
Solution Approach 1:
The patent applies local quality by dividing each unit area into multiple sub-areas and assigning different filter coefficients to each sub-area based on its local image features. This allows the filtering process to adapt to local variations in image characteristics while maintaining a structured approach that balances complexity and performance.
Solution Approach 2:
The patent segments each unit area into multiple sub-areas, allowing independent filter coefficient determination for each sub-area. This segmentation enables the system to capture local image features more effectively while maintaining computational feasibility through hierarchical processing.
2Manufacturing precision
If filter coefficients are changed according to target area position or target pixel position, then filtering appropriateness for different areas is improved, but device complexity increases
Solution Approach 1:
The patent implements dynamics by making filter coefficients variable based on target area position and target pixel position within the unit area. The filter coefficient changing means dynamically selects or calculates appropriate coefficients for each sub-area and pixel position, allowing the filtering process to adapt to local image features while maintaining a systematic approach to manage the complexity.
Data Source
AI summary
An image filter (100) for calculating a pixel value of target pixel in an output image from a pixel value of each of a pixel or pixels belonging to a filter area in an input image by using a filter coefficient vector V, is configured to include a filter coefficient vector changing section (120) for changing the filter coefficient vector V according to at least either where the target area is in the input image, or where the target pixel is in the output image.