Convolutional Neural Network Facial Feature Extraction via Segmentation
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Current face recognition technologies face challenges in efficiently generating and processing facial features from images, leading to high data requirements and memory occupation, which hinders processing efficiency.
Innovation Solution
A method and apparatus utilizing a first convolutional neural network to generate a feature area image set from a face image, followed by a second convolutional neural network to extract area facial features, allowing for information sharing and reducing data volume, thereby improving generation efficiency.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If conventional face recognition technology processes entire face images through neural networks, then recognition accuracy is maintained, but memory occupation and data requirements increase significantly
Solution Approach 1:
The face image is divided into multiple feature area images based on detected key points (eyes, nose, mouth, etc.). Each feature area image is processed independently through the neural network to generate corresponding area facial features. This segmentation approach reduces the amount of data that needs to be stored and processed while maintaining recognition accuracy, as only relevant feature regions are extracted and analyzed rather than the entire face image.
2Measurement precision
If conventional face recognition technology processes entire face images, then complete facial information is captured, but processing efficiency decreases
Solution Approach 1:
The method extracts only the essential feature areas (eyes, nose, mouth, chin, etc.) from the complete face image using key point detection. By taking out and processing only these critical feature regions rather than the entire face image, the system maintains facial information completeness needed for accurate recognition while significantly improving processing efficiency through reduced computational load and faster feature extraction.
3Measurement precision
If the entire face image is input into the neural network, then comprehensive feature extraction is achieved, but data volume and computational requirements increase
Solution Approach 1:
The face image is segmented into multiple feature area images based on detected key points. Each feature area image contains specific facial features (eyes, nose, mouth, etc.) that are processed independently through the neural network. This segmentation maintains feature extraction comprehensiveness by covering all critical facial regions while reducing overall data volume, as the system processes smaller, targeted regions rather than the complete high-resolution face image.
Solution Approach 2:
Different feature areas of the face are processed with specialized attention according to their local characteristics. Each feature area image (eyes, nose, mouth, etc.) is extracted and processed to capture its specific local features. This local quality approach ensures comprehensive feature extraction by treating each facial region with appropriate processing focus while reducing data volume by not processing irrelevant areas with the same computational intensity.
Data Source
AI summary
Methods and apparatus for generating a facial feature. A specific embodiment of the method includes: acquiring a to-be-recognized face image; inputting the to-be-recognized face image into a first convolutional neural network to generate a feature area image set of the to-be-recognized face image, the first convolutional neural network being used to extract a feature area image from a face image; inputting each feature area image in the feature area image set into a corresponding second convolutional neural network to generate an area facial feature of the feature area image, the second convolutional neural network being used to extract the area facial feature of the corresponding feature area image; and generating a facial feature set of the to-be-recognized face image based on the area facial feature of the each feature area image in the feature area image set.