Image Processing Apparatus Region-Based Enhancement
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Current image quality enhancement technologies are limited in improving cognitive image quality, especially in pattern regions with regularity and mixed texture regions, and require external databases, making them unsuitable for devices with memory constraints like TVs and mobile terminals.
Innovation Solution
An image processing apparatus and method that reduces a low-definition image to a predetermined ratio, extracts visual features, and repeatedly applies image quality enhancement processes, partitioning the image into regions and overlapping them to enhance image quality without an external database.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If external database methods are used to enhance image quality in pattern and texture regions, then image quality improvement is achieved, but memory consumption increases to 200 MB making it unsuitable for TVs and mobile terminals
Solution Approach 1:
The image is divided into multiple regions (pattern regions and texture regions) and processed differently. Pattern regions are enhanced using one method while texture regions use another method, allowing optimized memory usage for each region type without requiring a large external database for the entire image.
Solution Approach 2:
Different enhancement strategies are applied to different regions of the image. Pattern regions receive enhancement suitable for their regularity characteristics, while texture regions receive enhancement suitable for their high-frequency characteristics, improving overall image quality without uniformly consuming high memory resources.
2Quantity of substance
If super-resolution using very deep convolutional networks (VDSR) is used, then processing can be done without external database, but image quality enhancement is insufficient in pattern regions and mixed texture regions
Solution Approach 1:
The image processing is segmented into region-based enhancement and network-based enhancement. By dividing the image into pattern regions and texture regions, the system applies appropriate enhancement methods to each, overcoming the limitations of uniform VDSR processing while maintaining lower memory consumption than external database methods.
Solution Approach 2:
Instead of applying heavy external database processing to the entire image, the patent applies targeted enhancement only to specific regions that need it (pattern and texture regions), achieving sufficient image quality improvement with reduced computational and memory resources.
3Device complexity
If image quality enhancement is performed on the entire image at once, then processing is simple, but distortion occurs in pattern regions and mixed texture regions
Solution Approach 1:
The image is segmented into multiple regions (pattern regions, texture regions, and other regions) that are processed separately. This regional segmentation prevents distortion in pattern and texture regions while maintaining manageable processing complexity through systematic region-based handling.
Solution Approach 2:
The processing approach dynamically adapts to different region types within the image. Rather than applying a static uniform processing method, the system dynamically selects appropriate enhancement strategies for each region type, improving image quality without excessive complexity.
Data Source
AI summary
Disclosed is an image processing apparatus. The present image processing apparatus comprises: an input unit for inputting an image; and a processor for shrinking the inputted image to a predetermined ratio, extracting a visual feature from the shrunken image, performing an image quality enhancement process reflecting the extracted visual feature in the inputted image, repeatedly performing, for a predetermined number of times, the shrinking, the extracting, and the image quality enhancement process on the image that has undergone the image quality enhancement process. The present disclosure relates to an artificial intelligence (AI) system and an application thereof that simulate the functions of a human brain, such as recognition, judgment, etc., by using a machine learning algorithm such as deep learning, etc.