Face Recognition Model Mask Fusion via Feature Point Alignment
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Face recognition systems often fail when individuals wear masks, increasing the risk of COVID-19 transmission as people may remove masks for recognition, compromising safety.
Innovation Solution
A face recognition method that involves extracting face sample images, detecting feature points, adjusting mask images to match face feature points, fusing these images with face sample images, and training a model to recognize masked faces using OpenCV algorithms, enabling accurate face recognition even with masks.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Reliability
If face recognition systems are used at entry and exit points, then security and access control are improved, but face recognition fails when people wear masks
Solution Approach 1:
The system performs preliminary action by pre-processing face sample images to extract feature points and pre-processing mask images to align them with face feature points before fusion. This preparatory work ensures that when a masked face is recognized, the system has already established the proper alignment and feature correspondence, thereby maintaining recognition accuracy despite mask wearing.
Solution Approach 2:
The patent uses feature points as an intermediary element between the face image and mask image. By detecting feature points on both the face and mask, and using these points to guide the alignment and fusion process, the system creates a bridge that allows accurate recognition even when the mask obscures part of the face.
2Reliability
If people remove masks for face recognition, then recognition accuracy is improved, but COVID-19 transmission risk increases
Solution Approach 1:
The patent converts the harmful effect of mask wearing (obscuring facial features) into a beneficial process by using the mask itself as part of the recognition template. Through image fusion technology, the mask image is combined with the face image to create a new composite template that includes both face and mask features, thereby transforming the obstacle into an advantage.
3Adaptability or versatility
If mask images are fused with face sample images, then mask wearing face recognition is improved, but system complexity increases
Solution Approach 1:
The patent applies segmentation by dividing the complex image fusion process into distinct stages: face feature point detection, mask feature point detection, alignment based on feature points, and fusion. This segmentation of the processing workflow makes the complex task more manageable and implementable through modular computational steps.
Data Source
AI summary
A device and a method and a non-transitory readable storage medium, for face recognition are provided, the method comprise: extracting a face sample image from a predetermined face sample library and performing feature point detection to obtain multiple face feature points; obtaining multiple mask images; selecting first to fourth face feature points from the multiple face feature points; defining a distance between the first and second face feature point as a mask image height, and defining a distance between the third and fourth face feature point as a mask image width; adjusting a size of each mask image according to the mask image height and the mask image width; fusing each adjusted mask image with the face sample image to obtain multiple face mask images to save into the predetermined face sample library; training a face recognition model based on the predetermined face sample library for face recognition.