Image Compression Using Segmented Achromatic and Chromatic Block Processing
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing image processing techniques fail to achieve a small file size for highly compressed PDFs due to inefficient unification of character blocks and often result in noticeable image quality deterioration when merging black or white characters into gray.
Innovation Solution
An image processing apparatus that includes edge detection, connection pixel extraction, binary image generation, and separate compression methods for character and background images, using a larger number of achromatic target colors in monochrome mode to minimize image quality loss and achieve smaller file sizes.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Quantity of substance
If character blocks are unified based on representative color difference and distance, then the number of character blocks is reduced, but the file size of the highly compressed PDF does not become small enough and many character colors remain
Solution Approach 1:
The patent segments character blocks into two distinct categories: achromatic character blocks (black, white, gray) and chromatic character blocks (colored). This segmentation allows for different processing strategies to be applied to each category, enabling more effective unification of achromatic blocks while preserving necessary chromatic information, thereby reducing the overall number of character blocks and improving compression efficiency.
Solution Approach 2:
The patent applies different unification criteria and processing methods to different types of character blocks. Achromatic character blocks undergo aggressive unification based on position and size parameters, while chromatic character blocks are handled separately to maintain color fidelity. This local quality approach allows optimal compression for each category without compromising overall image quality.
2Quantity of substance
If black or white character blocks are unified into gray character blocks, then the number of character blocks is reduced, but image quality deterioration becomes noticeable
Solution Approach 1:
The patent separates achromatic character blocks into distinct categories (black, white, gray) and processes them differently. By identifying and maintaining pure black and white character blocks as separate categories, the patent avoids converting them to gray, thereby preventing noticeable image quality deterioration while still achieving unification within each category.
Solution Approach 2:
The patent modifies the unification parameters and criteria based on the specific characteristics of character blocks. By adjusting the threshold values and unification conditions for achromatic blocks, the patent enables effective reduction of black and white character blocks to gray while maintaining acceptable image quality, and preserves pure black and white blocks when quality would be compromised.
3Device complexity
If the same compression method is applied to all character blocks, then the processing is simple, but the compression efficiency is insufficient and file size remains large
Solution Approach 1:
The patent divides character blocks into achromatic and chromatic categories and applies different compression methods to each category. Achromatic character blocks use one compression approach that exploits their uniformity, while chromatic character blocks use another approach that preserves color information. This segmentation-based differential compression significantly improves compression efficiency and reduces file size.
Solution Approach 2:
The patent applies locally optimized compression strategies tailored to the specific characteristics of different character block types. By analyzing the properties of each category and applying the most suitable compression method locally, the patent achieves superior overall compression efficiency without requiring excessively complex global processing.
Data Source
AI summary
An edge image generating unit detects edges in an original image and generates an edge image from the edges. A connection pixel extracting unit extracts connection pixel sets in the edge image. A binary image generating unit classifies colors of the connection pixel sets into a predetermined number of achromatic target colors and a predetermined number of chromatic target colors if a color mode is set as color, and classifies the colors of the connection pixel sets into a predetermined number of achromatic target colors if the color mode is set as monochrome, where the number of the achromatic target colors set in the color mode of monochrome is larger than the number of the achromatic target colors in the color mode of color.