Image Down-Conversion Using Pixel Difference Sorting
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing image down-conversion methods often compromise the contrast and texture of high-resolution images, leading to blurry displays and difficulties in accurate focus adjustment.
Innovation Solution
An image processing apparatus and method that generates a down-converted image by analyzing pixel differences and selecting pixel pairs based on these differences, using an image analysis unit to calculate and sort pixel differences and an output image generation unit to adjust and select pixel pairs for the down-converted image, thereby maintaining the original image's contrast and texture.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Quantity of substance
If averaging filter or pixel decimation processing is applied to down-convert a high-resolution image, then the number of pixels is reduced to match display unit capacity, but the contrast and texture of the original image deteriorate causing blurry display
Solution Approach 1:
The patent changes the parameter selection criteria from simple averaging to difference-based selection. By calculating absolute differences between adjacent pixels and selecting pixels with smaller differences, the method preserves local contrast and texture while reducing pixel count, thus resolving the contradiction between quantity reduction and quality maintenance
Solution Approach 2:
The patent applies different processing strategies to different regions of the image based on local characteristics. By evaluating adjacent pixel differences locally and selecting representative pixels based on these local variations, the method maintains local image quality (contrast and texture) while achieving overall down-conversion
2Quantity of substance
If multiple pixels are averaged or decimated for down-conversion, then the data capacity is reduced, but the contrast deteriorates making accurate focus adjustment difficult
Solution Approach 1:
The patent changes the parameter for pixel selection from uniform averaging to difference-based selection. By using absolute difference values between adjacent pixels as the selection criterion, the method preserves contrast information necessary for accurate focus adjustment while reducing data capacity through selective pixel output
3Quantity of substance
If conventional down-conversion methods are used to reduce pixel count for display, then the display unit capacity is matched, but the texture information is lost
Solution Approach 1:
The patent changes the selection parameter from position-based or simple averaging to difference-based selection. By calculating absolute differences between adjacent pixels and selecting pixels with smaller differences, the method preserves texture information while reducing pixel count to match display capacity
Solution Approach 2:
The patent introduces an intermediary calculation step (absolute difference computation) between the original image and the down-converted output. This intermediary parameter helps identify representative pixels that preserve texture information while achieving the desired reduction in pixel count
Data Source
AI summary
Provided is an image analysis unit that receives an image to be a down-conversion processing target as an input and generate adjacent pixel difference information used to select a pixel to be output to a down-converted image, and an output image generation unit that selects a pixel pair to be an adjacent pixel used for the down-converted image on the basis of the adjacent pixel difference information and generate the down-converted image by executing an adjustment of the selected pixel pair. The image analysis unit calculates a difference of an adjacent pixel pair of a processing block constituent pixel, generates a sorting result of arranging the differences in descending order, and the output image generation unit selects the pixel pair to be used for the down-converted image in descending order of difference values on the basis of the sorting result.