Image Processing Circuit for Confidence-Based Regional Correction
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing image correction methods fail to account for the varying characteristics of objects within an image, leading to unevenly corrected images when uniform correction is applied.
Innovation Solution
An image processing system that utilizes a segmentation circuit to divide images into regions based on pixel classes using a trained neural network model, generating segmentation and confidence maps, and applies correction effects region-by-region with intensity adjustments based on confidence levels, employing circuits for denoising, color correction, and sharpening.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Productivity
If uniform correction is applied to the entire image, then the correction process is simple and fast, but the image quality is uneven and unsatisfactory
Solution Approach 1:
The image is divided into multiple regions based on object characteristics detected by the neural network model. Different correction parameters are applied to each region, allowing uniform processing speed while achieving non-uniform correction results that match the varying characteristics of different image regions.
Solution Approach 2:
The patent applies different correction quality levels to different regions of the image based on their characteristics. Regions with important objects receive higher correction quality with more sophisticated processing, while other regions use simpler correction methods, thus achieving overall image quality improvement without uniformly high computational cost.
2Manufacturing precision
If region-by-region correction with neural network segmentation is applied, then image quality is improved, but the device complexity and processing time increase
Solution Approach 1:
The neural network model performs preliminary segmentation and object detection before the correction process. This preliminary action identifies regions of interest and prepares correction masks, allowing subsequent correction circuits to operate more efficiently with pre-defined parameters rather than processing the entire image uniformly.
Solution Approach 2:
The patent applies full correction processing only to specific regions containing important objects, while other regions receive reduced or no correction. This partial action approach focuses computational resources where they are most needed, reducing overall processing complexity while maintaining high correction precision for critical areas.
3Manufacturing precision
If high correction intensity is applied to all regions, then correction effectiveness is maximized, but natural image appearance is lost
Solution Approach 1:
Different correction intensities are applied to different regions based on their characteristics. Regions containing important objects like faces receive higher correction intensity for enhanced detail, while other regions maintain lower intensity to preserve their natural appearance. This local differentiation resolves the contradiction between correction effectiveness and image naturalness.
Data Source
Figure 1
Figure 2
Figure 3
AI summary
Provided are an image processing circuit, a system-on-chip including the same, and a method of improving quality of a first image. The image processing circuit includes a tuning circuit configured to receive a segmentation map including pixel-by-pixel class inference information of the first image and a confidence map including confidence of the class inference information, determine classes of respective pixels of the first image, correction effects for each pixel of the image, and correction values indicating intensity of the correction effects based on the segmentation map and the confidence map, and generate a correction map based on the classes and the correction values of the respective pixels; and at least one correcting circuit configured to generate an enhanced image by applying correction effects according to the correction values to the respective pixels of the first image based on the correction map.