Face Image Processing via Feature Vector and 2D Code Generation
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Face recognition technology poses a risk to user privacy as it can expose personal information when face images are captured and processed, leading to potential misuse.
Innovation Solution
A face image processing method and device that extracts features from 2D face images to generate feature vectors and decompose 3D face images into base images and weighting factors, creating 2D codes that can be used as secure keys without exposing the original face information, utilizing a combination of image information capture, 3D morphable models, and 2D code generation units.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If face images are captured and processed for recognition, then face recognition accuracy is improved, but user privacy exposure risk increases
Solution Approach 1:
The patent extracts only the essential facial features (landmarks, distances, ratios) from the complete face image, separating the recognition-critical information from the privacy-sensitive visual data. This extraction process enables accurate recognition while eliminating the ability to reconstruct the original face image from the stored features.
Solution Approach 2:
The patent creates an abstract feature representation (a simplified copy) of the face image that contains recognition information but cannot be used to recreate the original image. This feature vector serves as a functional substitute that maintains recognition capability while removing privacy risks associated with storing actual face images.
2Measurement precision
If detailed face features are stored for accurate recognition, then recognition precision is improved, but the risk of information misuse increases
Solution Approach 1:
The patent segments the face image into discrete feature points (landmarks) and computes derived features (distances, ratios, angles) between them. This segmentation transforms continuous image data into discrete, non-reversible feature values that are precise for recognition but useless for reconstructing the original image or identifying the person visually.
Solution Approach 2:
The patent transforms spatial image data into mathematical parameters (coordinates, distances, ratios, angles) that represent facial geometry. This parameter transformation changes the form of information from visual to numerical, enabling precise computational comparison while making the data incomprehensible and unusable for visual identification or misuse.
3Reliability
If complete face images are used for recognition, then recognition reliability is improved, but the complexity of privacy protection measures increases
Solution Approach 1:
The patent performs preliminary processing of the face image by extracting features and computing feature vectors before any recognition or storage operations. This preliminary transformation into feature space enables subsequent recognition operations to be performed on the already-processed feature data, eliminating the need for additional privacy protection measures during storage and comparison.
Data Source
AI summary
A face image processing method is provided. The face processing method comprising: extracting a plurality of features from a primary two-dimensional face image to generate a first feature vector; decomposing a three-dimensional face image into a plurality of base face images and a plurality of weighting factors corresponding to the base face images; generating a first two-dimensional code according to the first feature vector; and generating a second two-dimensional code according to the weighting factors.