Face ROI Filtering for Multi-Frame Noise Reduction
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing image processing techniques that reduce noise by filtering multiple frames together often result in blurring of regions of interest, such as faces, compromising image quality.
Innovation Solution
Applying different amounts of filtering to regions of interest (ROIs) and remaining regions within frames, with weaker filtering on ROIs and stronger filtering on other regions, using threshold-based frame selection to determine the extent of blending.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Object-affected harmful factors
If filtering is applied across multiple frames to reduce noise, then noise reduction is improved, but blurring of regions of interest occurs
Solution Approach 1:
The patent applies different filtering strengths to different regions of the image. Specifically, weaker filtering is applied to regions of interest (such as faces detected via face detection) while stronger filtering is applied to non-ROI areas. This local differentiation allows noise reduction in background areas without blurring important subject details, directly resolving the contradiction between noise reduction and detail preservation.
2Object-affected harmful factors
If stronger filtering is applied to reduce noise, then noise reduction is improved, but image detail in critical areas deteriorates
Solution Approach 1:
The system dynamically adjusts filtering strength based on regional importance. Face detection identifies critical areas, and the filtering algorithm applies minimal filtering to these regions while applying stronger filtering to less important areas. This selective approach prevents information loss in critical regions while still achieving noise reduction overall.
Solution Approach 2:
The image is segmented into different regions based on importance: ROI (regions of interest) and non-ROI areas. Different filtering operations are applied to each segment, with weaker filtering for ROI and stronger filtering for non-ROI. This segmentation strategy allows differential noise reduction that preserves important details while removing noise from less critical areas.
Data Source
AI summary
An example device for image processing includes a memory and one or more processors coupled to the memory and configured to determine a region of interest (ROI) within a first frame and a second frame of a plurality of frames, apply a first amount of filtering on the ROI of the first frame and the second frame to generate a filtered ROI, apply a second amount of filtering on a remaining region of the first frame and the second frame to generate a filtered remaining region, combine the filtered ROI and the filtered remaining region to generate a filtered frame, and output the filtered frame.