FARE: A Deep Learning-Based Framework for Radar-based Face Recognition and Out-of-distribution Detection

2025-01-14

AI Summary

To address the dual challenges of contactless face recognition and out-of-distribution (OOD) anomaly detection, this paper proposes a dual-path deep learning framework leveraging 60 GHz short-range FMCW radar. Methodologically, it introduces two complementary input modalities—Range-Doppler and micro-Range-Doppler images—and designs a novel collaborative architecture comprising a primary classification pathway and multiple intermediate lightweight linear autoencoder pathways. A two-stage training strategy is adopted: first, triplet loss optimizes identity discrimination; then, the backbone is frozen while the autoencoder pathways are independently fine-tuned for OOD detection. Evaluated on a custom-built radar-based facial dataset, the system achieves 99.30% identity recognition accuracy and 96.91% AUROC for OOD detection—marking the first demonstration of high-accuracy joint modeling of identity recognition and OOD detection in the radar domain. This work establishes a new paradigm for privacy-preserving, seamless biometric authentication.

Technology Category

• Computer Vision: Biometrics, Face, Gesture & Pose
• Machine Learning: Multimodal Learning
• Intelligent Robots: Multimodal Perception & Sensor Fusion

Application Category

• Security and Privacy: Large-scale security measurements
• User Modeling, Personalization and Recommendation: Attacks and countermeasures in recommendation systems
• Systems and Infrastructure for Web, Mobile and WoT: Applied ML and AI for Web-based mobile applications

Abstract

In this work, we propose a novel pipeline for face recognition and out-of-distribution (OOD) detection using short-range FMCW radar. The proposed system utilizes Range-Doppler and micro Range-Doppler Images. The architecture features a primary path (PP) responsible for the classification of in-distribution (ID) faces, complemented by intermediate paths (IPs) dedicated to OOD detection. The network is trained in two stages: first, the PP is trained using triplet loss to optimize ID face classification. In the second stage, the PP is frozen, and the IPs-comprising simple linear autoencoder networks-are trained specifically for OOD detection. Using our dataset generated with a 60 GHz FMCW radar, our method achieves an ID classification accuracy of 99.30% and an OOD detection AUROC of 96.91%.

Problem

• Radar Technology
• Facial Recognition
• Anomaly Detection

Innovation

• FARE system
• Radar technology
• Phased training strategy