Lensless Image Restoration Region Selection for Secure Analysis
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing image capturing elements without lenses struggle to identify regions for restoration, leading to unauthorized disclosure and increased image analysis load due to uncontrolled restoration of entire images.
Innovation Solution
An image processing apparatus that identifies restoration regions using a restoration matrix, where output pixel values of multiple pixels have different angle-of-incidence directional sensitivities, allowing selective restoration and non-restoration of image regions.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If restoration is applied to the entire detection image, then image quality is improved, but information security deteriorates due to unauthorized disclosure
Solution Approach 1:
The detection image is divided into multiple restoration regions and non-restoration regions. The restoration process is applied selectively only to specific regions (e.g., regions containing objects of interest) while leaving other regions unrestored. This segmentation allows image quality improvement in necessary areas while maintaining information security in sensitive areas.
Solution Approach 2:
Different regions of the detection image are assigned different quality characteristics. Restoration processing is applied locally to specific regions where high image quality is needed, while other regions maintain their original unrestored state. This creates local quality differentiation that balances image utility with information protection.
2Measurement precision
If restoration is applied to the entire detection image, then image analysis accuracy is improved, but computational load increases
Solution Approach 1:
The image processing system segments the restoration task by applying restoration algorithms only to specific restoration regions rather than the entire detection image. This reduces the total computational load while maintaining analysis accuracy in the regions where restoration is applied.
Solution Approach 2:
Instead of applying restoration processing to the complete detection image (excessive action), the system applies restoration only to the necessary portions (partial action) - specifically to regions where improved image quality is needed for analysis. This reduces unnecessary computational expenditure while maintaining adequate analysis accuracy.
Data Source
AI summary
A restoration region where a restoration image is to be created by use of a restoration matrix is identified, the restoration region being in a region of a detection image obtained at an image capturing element that includes multiple pixels to receive incident beams that are incident thereon via neither an image capturing lens nor a pinhole and that is configured such that output pixel values of at least two pixels in the multiple pixels have mutually different characteristics in terms of angle-of-incidence directional sensitivities about incident beams from a subject. For example, the present disclosure can be applied to an image processing apparatus, an image capturing apparatus, an image capturing element, electronic equipment, a system, and the like.