Image Processing Apparatus Non-Integer Resizing RAW Data
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Conventional image processing systems are limited in resizing RAW data with non-integer magnification ratios, leading to image quality degradation due to pseudo color and moire issues when resizing pixel-thinned data.
Innovation Solution
An image processing apparatus and method that performs demosaicing followed by resizing with arbitrary magnification ratios, allowing conversion between RAW data formats with and without color interpolation, and then back to RAW data format, enabling flexible pixel adjustments without integer-based limitations.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Device complexity
If pixel thinning is used for resizing RAW data, then the processing complexity is reduced, but the magnification ratio is limited to 1/N and image quality deteriorates due to pseudo color and moire
Solution Approach 1:
The patent performs demosaicing processing before resizing to convert RAW data into full-color image data. This preliminary conversion enables subsequent resizing operations to use arbitrary magnification ratios without the limitations of pixel thinning, while maintaining image quality by avoiding pseudo color and moire artifacts.
Solution Approach 2:
The patent changes the data format parameter from RAW (single color component per pixel) to full-color image data (multiple color components per pixel) through demosaicing. This parameter change enables flexible magnification ratios in the resizing step while preserving image quality, resolving the contradiction between processing simplicity and magnification flexibility.
2Productivity
If pixel thinning is used for resizing, then the processing load is reduced, but image quality deteriorates due to pseudo color and moire artifacts
Solution Approach 1:
The patent performs demosaicing processing before resizing to convert RAW data into full-color image data. This preliminary conversion enables subsequent resizing operations to use arbitrary magnification ratios without the limitations of pixel thinning, while maintaining image quality by avoiding pseudo color and moire artifacts.
Solution Approach 2:
The patent introduces full-color image data as an intermediary format between RAW data and the final resized output. This intermediary format allows for high-quality resizing with arbitrary magnification ratios, avoiding the image quality deterioration that occurs when directly thinning RAW data pixels.
3Manufacturing precision
If demosaicing is performed before resizing with arbitrary magnification ratios, then image quality is maintained, but the processing complexity increases
Solution Approach 1:
The patent performs demosaicing processing before resizing to convert RAW data into full-color image data. This preliminary conversion enables subsequent resizing operations to use arbitrary magnification ratios without the limitations of pixel thinning, while maintaining image quality by avoiding pseudo color and moire artifacts.
Data Source
AI summary
Image data of a format in which pixel data having a value of one color component among a predetermined plurality of color components is regularly arranged is converted to a data format in which each item of pixel data has values of the plurality of color components. After performing resizing processing on this image data, the image data is returned to the original data format. Thus, it is possible to perform resizing with a magnification ratio that is not limited to one divided by an integer.