Facial Authentication with Virtual Mask Models
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Biometric access control systems face challenges with iris recognition due to longer detection times and performance degradation with glasses or contact lenses, and fingerprint recognition requires direct contact, while facial recognition systems struggle with masked faces, especially during the COVID-19 pandemic when mask-wearing is mandatory.
Innovation Solution
A user authentication device that detects face regions in masked or unmasked images, generates reference data from unmasked images, and authenticates users by comparing characteristics data from masked or unmasked images using main or secondary reference data, with the ability to generate virtual mask models to maintain accurate authentication.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Productivity
If facial recognition is used for access control, then authentication speed and user convenience are improved, but recognition performance deteriorates when the user wears a mask
Solution Approach 1:
The system performs preliminary actions by generating a 3D face model and virtual mask model during the registration phase (before authentication). This allows the system to pre-compute expected masked face characteristics, enabling fast and accurate authentication even when the user wears a mask during access control.
Solution Approach 2:
The system changes the parameter representation from 2D images to 3D face models. By converting face images into 3D models and generating corresponding virtual mask models, the system transforms the recognition problem into a 3D parameter comparison task, which maintains accuracy despite mask occlusion in 2D images.
2Measurement precision
If the system generates virtual mask models to maintain authentication accuracy with masked faces, then recognition performance is improved, but device complexity increases
Solution Approach 1:
The complex operation of generating 3D face models and virtual mask models is performed in advance during registration, not during authentication. This shifts the computational complexity to a one-time setup phase, keeping the real-time authentication process simple and efficient.
Solution Approach 2:
The system creates a virtual copy (virtual mask model) of the user's face in 3D space that mimics how their actual face appears when wearing a mask. This copying approach allows the system to compare real masked images against the pre-created virtual mask model, simplifying the authentication logic while maintaining high accuracy.
Data Source
AI summary
Embodiments relate to a user authentication device configured to detect a face region in a target object image including at least part of a face of a target object, recognize masked or unmasked in the face region, extract target object characteristics data from the face region of the target object image, call reference data and authenticate if the target object is a registered device user based on the called reference data and the target object characteristics data. The reference data is generated from an unmasked image of the registered device user.