Image Decoding Apparatus Chrominance Loop Filter Memory Optimization
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
In video coding schemes like Versatile Video Coding, processing for chrominance images is inadequate compared to luminance images in loop filters, and requires significant memory for filter processing.
Innovation Solution
An image decoding apparatus with filter units that apply filters to both luminance and chrominance images, using filter coefficients derived from coded data to limit pixel value changes within a prescribed range, adjusting this range based on slice type and bit depth for efficient processing.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Manufacturing precision
If loop filter processing for chrominance image is performed by referring to luminance image, then image quality of chrominance image is enhanced, but a large amount of memory is required to store intermediate images
Solution Approach 1:
The patent extracts and processes only the necessary filter coefficients from the luminance image, rather than storing and processing entire intermediate images. This selective extraction approach maintains image quality enhancement while significantly reducing memory requirements by focusing only on the essential filtering parameters.
Solution Approach 2:
The patent segments the filter processing into distinct stages: deriving filter coefficients from luminance image data, applying these coefficients to chrominance images, and limiting pixel value changes. This segmentation allows efficient memory usage by processing coefficient data separately from image data.
2Manufacturing precision
If processing for chrominance image is performed in loop filter, then image quality is enhanced, but processing efficiency is reduced compared to luminance image processing
Solution Approach 1:
The patent changes the processing parameters by deriving specific filter coefficients from luminance image data and applying them to chrominance images with controlled pixel value changes. This parameter-based approach enables efficient chrominance processing that maintains quality enhancement while improving processing speed through optimized coefficient application.
3Device complexity
If filter coefficient is derived from coded data, then device complexity is reduced, but filter processing accuracy may be compromised
Solution Approach 1:
The patent uses filter coefficients that are derived from and copy essential characteristics of the luminance image data. These coefficients capture the necessary filtering properties without requiring storage of the entire luminance image, thus reducing device complexity while maintaining filter processing accuracy through efficient data representation.
Data Source
AI summary
An image decoding apparatus includes: a first filter unit configured to apply a first filter to an image (a luminance image and a chrominance image); a second filter unit configured to apply a second filter to an output image of the first filter; a filter set derivation unit configured to decode a filter coefficient from coded data; and a third filter unit configured to apply a third filter to an output image of the second filter by using the filter coefficient. In a case that the third filter unit performs filter processing for the luminance image by using a luminance output image of the second filter, the third filter unit performs processing of clipping an amount of change of a pixel value through the filter processing within a prescribed range of value.