Face Image Validation via Keypoint Spatial Consistency
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing face recognition methods face challenges in accurately identifying faces with low-quality input images, including small face areas, extreme head poses, and environmental noise, leading to misrecognition and reduced accuracy.
Innovation Solution
A processor-implemented method that detects facial keypoints, determines face and facial feature areas, and filters invalid face images based on preset conditions such as shape and positional relationships, using a combination of facial keypoint detection and face classification models to enhance recognition accuracy.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If traditional face recognition methods are used on low-quality images, then processing speed is maintained, but recognition accuracy deteriorates due to small face areas, extreme head poses, and environmental noise
Solution Approach 1:
The patent applies preliminary action by performing face validity detection before main recognition processing. The system pre-processes images by detecting facial keypoints, determining face and facial feature areas, and validating their spatial relationships. This preliminary validation filters out invalid images (with small face areas, extreme poses, or noise) before they reach the main recognition model, improving overall accuracy while maintaining processing efficiency for valid images.
2Reliability
If strict validation conditions are applied to filter invalid images, then recognition accuracy is improved, but processing time increases due to additional detection and validation steps
Solution Approach 1:
The patent applies local quality by performing validation checks selectively on specific local features rather than uniformly processing all images. The system focuses validation efforts on critical local aspects: facial keypoint detection, face area determination, and spatial relationship verification between face and feature areas. This localized validation approach ensures reliability for images that need it while minimizing time loss on already-valid images.
3Measurement precision
If comprehensive spatial and shape conditions are checked for facial features, then false recognition is reduced, but computational load increases due to multiple condition evaluations
Solution Approach 1:
The patent applies preliminary action by pre-establishing spatial and shape condition criteria for facial features before processing images. The system pre-defines valid spatial relationships between face areas and facial feature areas, as well as acceptable shape characteristics. During image processing, these pre-defined conditions are efficiently evaluated rather than computed from scratch, reducing computational energy consumption while maintaining high detection precision.
Data Source
AI summary
A processor-implemented method with image processing includes detecting facial keypoints from an input face image determining a face area of the input face image and a facial feature area of the input face image based on the facial keypoints, and determining the input face image to be an invalid face image in response to the facial feature area satisfying a first preset condition, wherein the first preset condition comprises either one or both of a shape condition regarding a shape of the facial feature area, and a position condition regarding a relationship between a position of the facial feature area and a position of the face area.