Image Processing Device Two-Dimensional Distribution Equalization
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Image processing technologies face challenges in maintaining image quality due to non-uniform pixel data distribution when mixing an even number of pixels, leading to degradation in image resolution and the occurrence of false colors, especially during resizing processes for HD movies and small RAW data recording.
Innovation Solution
The implementation of a two-dimensional distribution equalization process that corrects pixel data distribution uniformly in both horizontal and vertical directions by dividing source image RAW data into luminance and color carrier data, performing interpolation and resizing independently for each, and re-synthesizing the data to ensure uniform distribution and prevent false colors.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Productivity
If pixel mixing reading method is used to reduce data size, then data volume is reduced and processing time is shortened, but pixel data distribution becomes non-uniform causing image quality degradation
Solution Approach 1:
The patent segments the pixel data processing into two independent one-dimensional distribution equalization processes (horizontal and vertical directions) rather than treating it as a single two-dimensional problem. This segmentation allows each direction to be processed independently, maintaining computational efficiency while achieving uniform pixel distribution and preventing image quality degradation.
Solution Approach 2:
The patent transforms the two-dimensional pixel distribution problem into two separate one-dimensional problems by processing horizontal and vertical directions independently. This dimensionality reduction simplifies the computational complexity while effectively achieving uniform distribution across the entire image plane, thus maintaining both processing speed and image quality.
2Adaptability or versatility
If resizing process is performed on non-uniformly distributed pixel data, then data size is changed for HD movies and small RAW recording, but false colors occur and image quality deteriorates
Solution Approach 1:
The patent performs distribution equalization as a preliminary step before the resizing process. By uniformly distributing pixel data in both horizontal and vertical directions before resizing, the patent prevents false color artifacts and maintains image quality during the subsequent resizing operation for HD movies and small RAW recording modes.
Solution Approach 2:
The patent changes the distribution parameter of pixel data from non-uniform to uniform through the distribution equalization process. This parameter transformation ensures that when resizing is subsequently applied, the pixel data maintains proper spatial relationships, preventing false colors and preserving image quality across different output resolutions.
3Device complexity
If conventional one-dimensional interlace conversion is used, then processing is simplified, but false colors occur and fine-resolution image quality is not ensured
Solution Approach 1:
The patent segments the distribution equalization into two independent one-dimensional processes (horizontal and vertical) that can be efficiently implemented without complex two-dimensional processing. This segmentation achieves fine-resolution image quality with uniform pixel distribution while keeping the processing complexity manageable through independent directional processing.
Data Source
AI summary
In a first filtering processing step, RAW data of a source image is subjected to a pixel-based filtering process along a first array direction to divide the RAW data into a first luminance data and a first color carrier data. In a first luminance distribution equalization processing step, the luminance distribution of the first luminance data in the first array direction is corrected to be uniform to produce a second luminance data. In a first color array reproduction processing step, the first color carrier data is re-synthesized with the second luminance data to produce a first multiple-color array data. In a second filtering processing step, the first multiple-color array data is subjected to a pixel-based filtering process along a second array direction to divide the first multiple-color array data into a third luminance data and a second color carrier data. In a second luminance distribution equalization processing step, the luminance distribution of the third luminance data in the second array direction is corrected to be uniform to produce a fourth luminance data. In a second color array reproduction processing step, the second color carrier data is re-synthesized with the fourth luminance data to produce a second multiple-color array data.