Block-Based Image Compression Using Edge-Aware Frequency Selection
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing image processing devices face challenges in achieving a high compression rate while maintaining image quality, particularly in distinguishing between different types of image data such as text and natural images, which affects memory storage and transmission efficiency.
Innovation Solution
An image processing device that divides input image data into blocks, determines the type of each block based on edge elements, converts color components into frequency components, and compresses the data by identifying and suppressing unnecessary frequency components specific to each type, using techniques like discrete cosine transform (DCT) and pattern matching for text images.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Reliability
If lossless compression processing is used for text images, then image quality is maintained, but compression rate is limited
Solution Approach 1:
The patent applies different compression strategies to different regions of the image based on local characteristics. Text regions undergo lossless compression to maintain sharp edges and character definitions, while natural image regions use lossy compression to achieve higher compression rates. This local differentiation resolves the contradiction by allowing high compression rates overall while maintaining quality where it matters most.
Solution Approach 2:
The image is divided into multiple blocks, and each block is independently analyzed and compressed according to its specific characteristics. This segmentation allows the system to identify text blocks versus natural image blocks and apply appropriate compression methods to each, thereby achieving both high compression rates and maintained quality in text regions.
2Productivity
If lossy compression processing is used for natural images, then compression rate increases, but image quality deteriorates
Solution Approach 1:
Lossy compression is applied selectively only to natural image regions where quality degradation is less perceptible, while text regions are protected with lossless compression. This resolves the contradiction by confining quality deterioration to areas where it matters least while maximizing compression rates in those same areas.
Solution Approach 2:
By segmenting the image into text and natural image blocks, the patent can apply lossy compression only to the natural image portions, achieving high overall compression rates without significantly impacting perceived image quality since text regions maintain their sharpness and definition.
3Device complexity
If image data is processed as a whole, then processing is simple, but compression rate is reduced due to inability to differentiate image types
Solution Approach 1:
The patent segments the image into multiple blocks and processes each block independently after identifying its type. This segmentation enables differentiation between text and natural image regions, allowing the system to achieve high compression rates by applying appropriate methods to each segment while keeping the overall processing framework relatively simple and modular.
4Measurement precision
If edge detection is performed for every block to determine image type, then compression accuracy improves, but processing time increases
Solution Approach 1:
The patent performs edge detection and image type determination only for necessary blocks rather than uniformly processing every block with the same level of analysis. This partial action approach maintains high compression accuracy by identifying image types where it matters most while reducing processing time by avoiding redundant analysis in already-identified regions.
Data Source
AI summary
An image processing device that processes input image data includes: an image dividing unit that divides the input image data into a plurality of blocks; an image type determining unit that determines the type of the input image data for every block on the basis of the number of edge elements included in the block; a frequency component conversion unit that converts color components of the input image data into frequency components for every block; and an image compression unit that compresses corresponding image data for every block by specifying unnecessary components of the converted frequency components of the image data on the basis of the determined type of the image data and suppressing or removing the specified unnecessary components.