Dynamic Facial Motion Recognition for Biometric Binding
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Conventional face recognition systems are unreliable due to mistaken recognition of similar individuals or photographs, and accidental unlocking, as they rely solely on static facial images rather than dynamic facial features.
Innovation Solution
A face recognition system that captures a video clip of a user's facial expressions over a time frame, extracts and compares facial feature variations using a processing unit, and performs a binding operation only if the deviation falls within a threshold, enhancing security and reliability through dynamic face recognition.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Reliability
If static facial image recognition is used, then the system is simple and fast to operate, but the reliability is low due to mistaken recognition of similar individuals or photographs
Solution Approach 1:
The patent transitions from static facial image recognition to dynamic facial motion recognition. The system captures video clips and analyzes temporal variations in facial features, transforming the recognition process from a static snapshot to a dynamic temporal analysis, thereby improving reliability while managing complexity through structured processing
Solution Approach 2:
The patent adds the time dimension to facial recognition by analyzing facial feature variations over time. Instead of comparing only spatial coordinates of facial landmarks in a single image, the system incorporates temporal sequences of facial expressions and movements, creating a multi-dimensional recognition space that distinguishes between identical individuals and photographs
2Reliability
If static facial image recognition is used, then the operation is simple and quick, but false identification occurs due to similar appearances or photographs
Solution Approach 1:
The system performs preliminary actions by capturing a sequence of facial images and extracting facial feature points before the actual recognition comparison. This preparation includes detecting facial landmarks, tracking their movements, and computing variation sequences in advance, which streamlines the subsequent authentication process while maintaining high accuracy
Solution Approach 2:
The system uses dynamic facial motion analysis to rapidly distinguish between real users and photographs or look-alikes. By analyzing the temporal patterns of facial feature movements during a short video clip, the system achieves quick and accurate identification without requiring lengthy verification processes
3Reliability
If static facial image recognition is used, then the system responds quickly, but accidental unlocking occurs due to mere appearance in front of the device
Solution Approach 1:
The system requires dynamic facial motions and expressions rather than a static pose. Users must perform specific facial movements or expressions during the recognition process, which prevents accidental unlocking from mere appearance while still maintaining ease of operation through intuitive facial gestures
Solution Approach 2:
The system implements preliminary verification by analyzing facial feature variations over time before granting access. This preliminary check ensures that the person present is performing the correct facial motions, preventing unauthorized access while keeping the process simple for legitimate users
Data Source
AI summary
A face recognition system based upon a facial motion including an image capturing device, a storage device and a processing unit. The storage device stores facial information of an intended user. The facial information contains data of the intended user's facial feature and is associated with a binding operation. The image capturing device captures a facial video clip where a user makes a series of facial expressions over a time frame. The facial video clip of the user contains image frames. The processing unit extracts at least one facial feature of the user from the image frames and calculates the variation of that over the time frame. The processing unit compares the variation of the facial feature of the user with the facial information of the intended user. If the deviation between them falls within a threshold, the processing units goes on to perform the binding operating.