Continuous Face Verification Enrollment via Off-Pose Image Buffering
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Conventional face verification systems require users to enroll and authenticate in a strictly frontal pose, leading to cumbersome enrollment processes and potential authentication failures when users are not square with the camera, as the templates lack data for off-pose positions.
Innovation Solution
Implementing a continuous enrollment method where a computing device captures a series of images during authentication, buffering off-pose positions and adding them to the user's face template once authentication is successful, allowing for immediate authentication without requiring users to turn towards the camera.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Reliability
If multiple facial poses are captured during enrollment, then authentication accuracy for off-pose positions is improved, but the enrollment process becomes cumbersome and complex
Solution Approach 1:
The system performs preliminary action by capturing and buffering a series of images during the authentication process itself, before the face template is updated. This allows off-pose positions to be collected naturally during authentication rather than requiring a separate complex enrollment process, resolving the contradiction by preparing data in advance without increasing enrollment complexity
Solution Approach 2:
The system uses self-service by automatically capturing, buffering, and updating face template data during the authentication process without requiring explicit user enrollment actions. The user simply needs to present their face for authentication, and the system autonomously collects off-pose positions and updates the template, eliminating the need for cumbersome manual enrollment procedures
2Reliability
If multiple facial poses are captured during enrollment, then authentication for off-pose positions is improved, but the enrollment process becomes time-consuming
Solution Approach 1:
The system merges the enrollment function into the authentication process by capturing images and updating the face template during authentication. This combines two separate processes into one, allowing off-pose position collection to occur during the brief authentication moment rather than requiring a separate time-consuming enrollment session
Solution Approach 2:
The system implements continuity of useful action by continuously capturing images at multiple time points during authentication and continuously updating the face template with new off-pose positions. This ongoing process ensures the template is constantly improved without interrupting the authentication flow or requiring separate enrollment time
3Productivity
If the face template is updated with off-pose images, then authentication speed for off-pose positions is improved, but the system complexity increases
Solution Approach 1:
The system uses feedback by evaluating each captured image against the current face template during authentication, determining whether the image represents a useful off-pose position, and then updating the template accordingly. This feedback loop ensures the template is improved with high-quality data while maintaining system simplicity through automated decision-making
Solution Approach 2:
The system applies dynamics by making the face template updateable and adaptable over time. The template evolves from a static frontal-view-only representation to a dynamic structure that continuously incorporates new off-pose positions captured during authentication, allowing the system to adapt to actual user behavior patterns
Data Source
AI summary
Techniques for performing continuous enrollment for face verification are provided. In one embodiment, a computing device can receive, from a user, an indication that the user wishes to authenticate himself/herself with the computing device via face verification. In response to the indication, the computing device can capture, using a camera, a series of images of the user's face and can authenticate the user by evaluating each of the series of images against a face template for the user, where the user is authenticated based on an N-th image in the series. Once the user has been authenticated, the computing device can select one or more images from the series prior to the N-th image and can add the selected images to the user's face template.