Latent-space metrics for Complex-Valued VAE out-of-distribution detection under radar clutter

TL;DR

This paper explores the use of complex-valued Variational AutoEncoders and various detection metrics for radar out-of-distribution detection, comparing their effectiveness against classical methods on synthetic and real radar data.

Contribution

It introduces and evaluates new complex-valued VAE-based detection metrics for radar OOD detection, providing a comparative analysis with classical detectors.

Findings

CVAE-based metrics outperform classical detectors in certain scenarios

Latent-space scores provide effective OOD detection signals

Performance varies between synthetic and experimental radar data

Abstract

We investigate complex-valued Variational AutoEncoders (CVAE) for radar Out-Of-Distribution (OOD) detection in complex radar environments. We proposed several detection metrics: the reconstruction error of CVAE (CVAE-MSE), the latent-based scores (Mahalanobis, Kullback-Leibler divergence (KLD)), and compared their performance against the classical ANMF-Tyler detector (ANMF-FP). The performance of all these detectors is analyzed on synthetic and experimental radar data, showing the advantages and the weaknesses of each detector.