Image Reconstruction via Adaptive Regularization Strength Estimation
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing image processing technologies face difficulties in allowing non-specialist users to appropriately adjust regularization strengths for image restoration, leading to suboptimal image quality due to the need for precise parameter setting.
Innovation Solution
An information processing device that calculates variation amounts between pixel values and peripheral pixels, estimates regularization strengths based on attribute reliability and image quality information, and generates reconstructed images using these strengths.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Manufacturing precision
If regularization strength parameters are manually adjusted by users to achieve optimal image restoration quality, then image quality improves, but ease of operation deteriorates because non-specialist users cannot appropriately adjust parameters
Solution Approach 1:
The system automatically calculates and determines appropriate regularization strength parameters by analyzing the input image's pixel variation characteristics, eliminating the need for manual user adjustment. The image itself provides the information needed to set optimal parameters through automated analysis of variation amounts between peripheral pixels.
Solution Approach 2:
The system dynamically determines regularization strength parameters based on the actual image content by calculating pixel variation amounts. Instead of using fixed or manually-set parameters, the system adapts parameters automatically according to the local characteristics of each region in the image.
2Manufacturing precision
If multiple parameter values are required for image processing (variation amounts and regularization strengths), then manufacturing precision improves, but device complexity increases due to multiple parameters needing specification
Solution Approach 1:
The system extracts only the essential variation amount information from the input image by analyzing pixel differences between peripheral pixels. This extracted variation amount is then used to derive the regularization strength, eliminating the need for users to specify multiple independent parameters.
Solution Approach 2:
The system combines the specification of variation amounts and regularization strengths into a single automated process. By calculating variation amounts from the image and automatically deriving regularization strengths, the system merges what would otherwise be separate parameter specifications into one unified automatic determination.
Data Source
AI summary
An information processing device acquiring to the present invention includes: a CPU; and a memory storing a program, wherein the CPU, by the program, configures: a variation-amount calculating unit that, for an input image, calculates a variation amount between a value of a predetermined pixel of the input image and values of peripheral pixels of the predetermined pixel; an attribute reliability unit that, based on an attribute that is a property of pixels within a specified area in the input image and the variation amount, calculates attribute reliability of the pixel; a regularization strength estimating unit that, based on image quality information about image quality for the attribute and the attribute reliability, estimates a regularization strength of the pixel; and an image reconstructing unit generates a reconstructed image that is an image acquired by reconstructing the input image by using the regularization strength.