Fingerprint Anti-Counterfeiting Network for Cross-Sensor Recognition
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing fingerprint recognition technologies face challenges in achieving high recognition accuracy due to environmental factors and the need for separate neural network models for different sensor types, limiting their applicability and resource efficiency.
Innovation Solution
A method and apparatus for training a fingerprint anti-counterfeiting neural network that uses a classification network with preprocessing, fusion, and classification subnetworks to process data from multiple sensor types, incorporating a loss function to iteratively refine the network, enabling it to recognize fingerprints across various sensors and reduce environmental impact.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If separate neural network models are trained for different sensor types, then recognition accuracy for each sensor type is improved, but device complexity and training resource consumption increase
Solution Approach 1:
The patent implements a universal neural network model that can process fingerprint data from multiple sensor types (capacitive, optical, ultrasonic) simultaneously. The model uses a unified architecture with sensor-type identification and adaptive processing pathways, eliminating the need for separate models for each sensor type while maintaining high recognition accuracy across all sensor varieties.
Solution Approach 2:
The neural network model is segmented into distinct processing pathways for different sensor types, with each pathway optimized for its specific sensor characteristics. The model includes separate feature extraction modules for capacitive, optical, and ultrasonic sensors, allowing specialized processing within a unified framework that reduces overall complexity compared to maintaining entirely separate models.
2Productivity
If fingerprint data is collected directly by sensors in various environments, then data collection efficiency is improved, but training effectiveness deteriorates due to environmental factors like temperature fluctuation and film interference
Solution Approach 1:
The patent extracts and removes environmental interference factors from the fingerprint data processing pipeline. It implements preprocessing modules that specifically target and eliminate effects from temperature fluctuations, film interference, and other environmental conditions before the data reaches the neural network, thereby maintaining high training effectiveness despite continued efficient data collection in various environments.
Solution Approach 2:
The patent introduces intermediary processing layers between the sensor and the neural network that act as mediators to compensate for environmental effects. These intermediary modules include temperature compensation algorithms and film interference correction mechanisms that adjust the raw sensor data to account for environmental conditions, preserving both collection efficiency and training reliability.
3Measurement precision
If a trained neural network model is used for one sensor type, then recognition accuracy for that sensor type is improved, but adaptability to other sensor types is lost
Solution Approach 1:
The patent creates a multi-functional neural network model that maintains high recognition accuracy across multiple sensor types simultaneously. The unified model incorporates sensor type identification and adaptive processing that allows it to achieve sensor-specific performance without requiring separate trained models, thereby achieving both accuracy and adaptability.
Solution Approach 2:
The neural network model employs dynamic adaptation mechanisms that allow it to adjust its processing parameters and pathways based on the detected sensor type. This dynamic reconfiguration enables the model to optimize its performance for each sensor type in real-time while maintaining a single trained model structure, achieving both specialized accuracy and broad adaptability.
Data Source
AI summary
A method for training a fingerprint anti-counterfeiting neural network, a method for fingerprint anti-counterfeiting and an apparatus are provided. The method includes: obtaining groups of training data, each group of the training data includes: first raw domain data, second raw domain data, and third raw domain data; and training an initial classification network using the plurality of groups of training data to obtain a target classification network. The initial classification network includes a fusion subnetwork and a classification subnetwork The fusion subnetwork generates a first fingerprint matching pair based on a feature description matrix of the first raw domain data and the second raw domain data, and generates a second fingerprint matching pair based on a feature description matrix of the first raw domain data and the third raw domain data. The classification subnetwork performs fingerprint classification and recognition based on the first and second fingerprint matching pairs.