Joint Spoof-Detection Training With Performance-Weighted Data
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing spoof-detection systems require time-consuming and manually intensive training processes, and are not efficient in identifying various types of biometric spoofing attacks, leading to vulnerabilities in user authentication systems.
Innovation Solution
A method for jointly training multiple spoof-detection systems using a single data set, where each system is weighted based on its performance to improve accuracy and efficiency, allowing simultaneous training across different types of biometric data.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Reliability
If multiple spoof-detection systems are trained separately with pre-selected training data, then each system can be optimized for specific spoof types, but the training process becomes time-consuming and requires substantial manual work
Solution Approach 1:
The patent combines multiple spoof-detection systems into a single joint training framework where multiple detection models are trained simultaneously using a unified loss function and shared training data, eliminating the need for separate training processes while maintaining detection accuracy across different spoof types
Solution Approach 2:
The training system is designed to handle multiple types of biometric data (fingerprints, faces, iris) and multiple spoof types simultaneously through a universal training framework that adapts to different detection scenarios without requiring separate specialized training processes
2Measurement precision
If training data is pre-selected and sorted for each spoof-detection system, then training can be more targeted and effective, but substantial manual work is involved
Solution Approach 1:
The system automatically selects and weights training data samples based on their relevance to different spoof types and detection systems, using algorithms to determine optimal sample selection and weighting without requiring manual intervention, thereby maintaining training effectiveness while eliminating manual work
3Adaptability or versatility
If separate training is used for different spoof-detection systems, then each system can be specialized, but resource requirements and complexity increase
Solution Approach 1:
Multiple spoof-detection systems are merged into a single joint training architecture where models share common components and training resources, reducing overall system complexity while maintaining the ability to detect multiple spoof types through specialized detection heads or layers
Data Source
Figure 1
Figure 2
Figure 3
AI summary
A computer-implemented method for jointly training at least two different spoof-detection systems using a single data set, the single data set comprising at least two sub-sets of training data, wherein each sub-set comprises at least two different kinds of training data, the method comprising: - training, in a first training cycle, the at least two different spoof-detection systems a first sub-set of training data - determining, based on a result of the training, for each of the at least two different spoof-detection systems, a score for each of the at least two different kinds of training data in the sub-set of training data - weighting, based on the scores, each kind of training data in a second sub-set of training data to obtain weighted training data - training, in a subsequent training cycle, the at least two different spoof-detection systems using the weighted training data.