Neural Network-Powered Finger-Drawn Biometric Authentication

2025-11-14

AI Summary

This work addresses the need for convenient and secure authentication on mobile devices by proposing a lightweight touchscreen-based biometric authentication method using finger-drawn digits. It introduces, for the first time, a novel biometric modality that jointly models both dynamic (e.g., stroke timing, velocity) and static (e.g., shape, topology) behavioral characteristics of user-drawn digits on touchscreens. Methodologically, it integrates an improved Inception-V1 architecture with a custom lightweight CNN for user identification, and designs a convolutional–fully connected hybrid autoencoder for anomaly detection and liveness verification. Evaluated on a self-collected real-world finger-drawing dataset, the classification model achieves 89% authentication accuracy—while reducing CNN parameters by 42%—and the autoencoder attains 75% anomaly detection accuracy. The framework supports on-device deployment and multi-layer security integration, delivering high security, low latency, and seamless user experience—establishing a practical, deployable paradigm for frictionless mobile authentication.

Technology Category

• Computer Vision: Biometrics, Face, Gesture & Pose
• Machine Learning: Calibration & Uncertainty Quantification
• Humans and AI: Interaction Techniques and Devices

Application Category

• User Modeling, Personalization and Recommendation: On-Device user modeling, personalization, and recommendation
• Security and Privacy: Large-scale security measurements
• Systems and Infrastructure for Web, Mobile and WoT: Applied ML and AI for Web-based mobile applications

Abstract

This paper investigates neural network-based biometric authentication using finger-drawn digits on touchscreen devices. We evaluated CNN and autoencoder architectures for user authentication through simple digit patterns (0-9) traced with finger input. Twenty participants contributed 2,000 finger-drawn digits each on personal touchscreen devices. We compared two CNN architectures: a modified Inception-V1 network and a lightweight shallow CNN for mobile environments. Additionally, we examined Convolutional and Fully Connected autoencoders for anomaly detection. Both CNN architectures achieved ~89% authentication accuracy, with the shallow CNN requiring fewer parameters. Autoencoder approaches achieved ~75% accuracy. The results demonstrate that finger-drawn symbol authentication provides a viable, secure, and user-friendly biometric solution for touchscreen devices. This approach can be integrated with existing pattern-based authentication methods to create multi-layered security systems for mobile applications.

Problem

• Developing neural network authentication using finger-drawn digits on touchscreens
• Comparing CNN and autoencoder architectures for biometric user verification
• Creating secure mobile authentication through simple finger-drawn digit patterns

Innovation

• CNN and autoencoder architectures for authentication
• Modified Inception-V1 and lightweight shallow CNN
• Finger-drawn digit patterns for biometric security