Near out-of-distribution detection for low-resolution radar micro-Doppler signatures (2022-05-12T00:00:00.000000Z)

TL;DR

An OODD use case for radar targets detection extensible to other kinds of sensors and detection scenarios is put forward, and a comparison of deep and non-deep OodD methods on simulated low-resolution pulse radar micro-Doppler signatures is proposed.

Abstract

Near out-of-distribution detection (OODD) aims at discriminating semantically similar data points without the supervision required for classification. This paper puts forward an OODD use case for radar targets detection extensible to other kinds of sensors and detection scenarios. We emphasize the relevance of OODD and its specific supervision requirements for the detection of a multimodal, diverse targets class among other similar radar targets and clutter in real-life critical systems. We propose a comparison of deep and non-deep OODD methods on simulated low-resolution pulse radar micro-Doppler signatures, considering both a spectral and a covariance matrix input representation. The covariance representation aims at estimating whether dedicated second-order processing is appropriate to discriminate signatures. The potential contributions of labeled anomalies in training, self-supervised learning, contrastive learning insights and innovative training losses are discussed, and the impact of training set contamination caused by mislabelling is investigated.

Authors

Martin Bauw

1 papers

S. Velasco-Forero

1 papers

J. Angulo

1 papers

TL;DR

Abstract

Authors

References42 items

ICLR 2021 Challenge for Computational Geometry & Topology: Design and Results

A Simple Fix to Mahalanobis Distance for Improving Near-OOD Detection

Deep Random Projection Outlyingness for Unsupervised Anomaly Detection

Micro-Doppler Signal Representation for Drone Classification by Deep Learning

MedMNIST Classification Decathlon: A Lightweight AutoML Benchmark for Medical Image Analysis

From unsupervised to semi-supervised anomaly detection methods for HRRP targets

Modeling the Distribution of Normal Data in Pre-Trained Deep Features for Anomaly Detection

Understanding Contrastive Representation Learning through Alignment and Uniformity on the Hypersphere

Matlab®

Differentiating through the Fréchet Mean

Simple and Effective Prevention of Mode Collapse in Deep One-Class Classification

Deep Multi-sphere Support Vector Data Description

Riemannian batch normalization for SPD neural networks

A Hermitian Positive Definite neural network for micro-Doppler complex covariance processing

Target Detection and Classification of Small Drones by Deep Learning on Radar Micro-Doppler

Second-Order Networks in PyTorch

Deep Semi-Supervised Anomaly Detection

Deep Anomaly Detection with Outlier Exposure

Deep One-Class Classification

Temporal Deep Learning for Drone Micro-Doppler Classification

geomstats: a Python Package for Riemannian Geometry in Machine Learning

Second-order Convolutional Neural Networks

Geometric Deep Learning: Going beyond Euclidean data

A Riemannian Network for SPD Matrix Learning

Barycentric subspace analysis on manifolds

Unsupervised Learning of 3D Structure from Images

Robust RX anomaly detector without covariance matrix estimation

Support Vector Data Description

Manifold Valued Statistics, Exact Principal Geodesic Analysis and the Effect of Linear Approximations

Anomaly detection: A survey

Isolation Forest

Classification with a Reject Option using a Hinge Loss

A Classification Framework for Anomaly Detection

Applied Multivariate Statistical Analysis

Estimating the Support of a High-Dimensional Distribution

LOF: identifying density-based local outliers

General notions of statistical depth function

Breakdown Properties of Location Estimates Based on Halfspace Depth and Projected Outlyingness

RADAR SIGNALS

What is projection pursuit

Ieee Transactions on Signal Processing Riemannian Medians and Means with Applications to Radar Signal Processing

Near OODD for low-resolution radar micro-Doppler signatures HAL PREPRINT

Field of Study