Perceptually Lossless Color Compression via Layer Segmentation
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Conventional compression methods in digital imaging either result in large file sizes with lossless compression or significant distortion with lossy compression, failing to achieve high compression rates with no perceptual loss, which is essential for applications like digital copiers, MFPs, and medical imaging.
Innovation Solution
The system employs computer vision-based methods for image segmentation and layer-based compression, inverting the image generation process to achieve high compression rates with minimized loss by accurately separating image layers and using region-specific coding, similar to human visual processing.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Reliability
If lossless compression is used, then image quality is preserved, but file size becomes very large
Solution Approach 1:
The patent segments the image into multiple layers including a mask layer, a foreground layer, and a background layer. Each layer is compressed separately using appropriate methods, allowing the mask layer to maintain high resolution while the color layers use lower resolution compression, achieving high compression ratios without perceptual loss.
Solution Approach 2:
Different layers are assigned different compression qualities based on their importance. The mask layer, which contains critical text and line art information, is encoded at full resolution, while the foreground and background layers are encoded at reduced resolution, optimizing the balance between compression and quality.
2Quantity of substance
If lossy compression is used, then file size is reduced, but distortion increases
Solution Approach 1:
By dividing the image into semantically distinct layers, the patent applies lossy compression only to layers where it is acceptable (foreground and background), while preserving the mask layer at full quality. This segmented approach enables high compression ratios without significant perceptual distortion in the overall image.
Solution Approach 2:
Instead of applying compression uniformly to the entire image and then trying to recover quality, the patent inverts the approach by first separating the image into layers and then applying compression selectively. This inversion of the compression process allows for high compression ratios while maintaining perceptual quality through intelligent layer-based processing.
3Measurement precision
If mask layer is encoded at full resolution, then text and line art information is preserved, but byte cost increases
Solution Approach 1:
The patent separates the mask layer from the color layers, allowing the mask layer to be encoded at full resolution to preserve text and line art accuracy, while the color layers are encoded at reduced resolution. This segmentation isolates the high byte cost to only the necessary mask layer, optimizing the overall compression efficiency.
Solution Approach 2:
The mask layer is assigned high quality encoding at full resolution because it contains critical text and line art information that must be preserved accurately. The foreground and background layers are assigned lower quality encoding at reduced resolution, as color information is less critical for document accuracy. This local quality assignment optimizes the balance between text accuracy and byte cost.
Data Source
AI summary
A system and method of color data compression may perform variations of MRC compression including taking into consideration means and/or variances within k×k cells of an image, background texture, background smoothed color, and gradient components for determination of which parts of an image belong to the background and foreground MRC layers, for calculating thresholds values for such determinations, and determining correctness of determined thresholds and polarity, and may determine the background and foreground layer colors based on the variances, the gradient components, and neighboring pixels of non-color-assigned pixels of the background and foreground layers.