Image Pixel Positioning for Compression Artifact Reduction
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Current image compression methods, such as those used for HDTV and higher resolutions, often result in loss of high-frequency information and visual artifacts like blockiness and pixelization, especially when applied to larger raster sizes, due to the use of two-dimensional low-pass filters and predictive loops that do not account for the human visual system.
Innovation Solution
A method that subdivides image pixels into blocks, calculates cumulative sums of differences, and allocates pixel positions to minimize these differences, using filtering intensity based on cumulative sums and probability calculations to position pixels, thereby reducing bit usage and preserving visual quality.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Productivity
If two-dimensional low-pass filter is applied to compress image signal, then compression is achieved, but high frequency information is lost and visual quality deteriorates
Solution Approach 1:
The image is divided into multiple blocks, and each block is further divided into sub-blocks for independent processing. This segmentation allows the algorithm to preserve high-frequency information in different regions while achieving compression, avoiding the uniform blurring effect of traditional low-pass filters.
Solution Approach 2:
The patent changes the compression approach from frequency-domain filtering to spatial-domain block-based parameter optimization. By adjusting block size, sub-block configuration, and positioning parameters, the system achieves compression while preserving visual quality without losing high-frequency information.
2Reliability
If conventional compression is applied to 4k2k raster, then video part remains acceptable, but graphics and colored texts are degraded
Solution Approach 1:
Different processing strategies are applied to different regions within blocks. The algorithm identifies and preserves high-frequency characteristics in text and graphic regions while allowing more compression in video regions, achieving local quality optimization that maintains both video acceptance and text/graphic sharpness.
Solution Approach 2:
The compression parameters are dynamically adjusted based on local image characteristics. The system analyzes each block's content type (video, text, graphic) and adapts the compression strength accordingly, enabling variable quality preservation across different image regions.
3Productivity
If predictive loops with SAD or MSE criteria are used, then compression efficiency is optimized, but visual artifacts appear due to mathematically equivalent but visually different solutions
Solution Approach 1:
The patent introduces visual quality feedback into the compression process. Instead of relying solely on mathematical error criteria (SAD/MSE), the system incorporates human visual system characteristics to evaluate and select positioning solutions, preventing the selection of mathematically optimal but visually degraded solutions.
Solution Approach 2:
The evaluation criterion is changed from pure mathematical error metrics to a hybrid metric that combines compression efficiency with visual quality assessment. This parameter change allows the system to avoid visual artifacts while maintaining good compression performance.
Data Source
AI summary
An image processing method, implemented in a calculator, includes applying a process by group of pixels to an original image. For each group of pixels, calculating a cumulative sum of value or position differences of the pixels of the group of pixels, and for each group of pixels, allocating in a final image signal a pixel position of the group of pixels to each pixel value of the group of pixels so as to minimize the cumulative sum of differences calculated for the group of pixels according to the differences calculated. For each group of pixels, determining a filtering intensity according to the cumulative sum of differences calculated for the group of pixels, and applying to the group of pixels a filtering having the filtering intensity.