Multi-Channel Replay Speech Detection using Acoustic Maps

TL;DR

This paper introduces acoustic maps as a novel spatial feature for replay speech detection, leveraging multi-channel recordings and a lightweight neural network to improve security in voice verification systems.

Contribution

The work proposes a new acoustic map feature derived from beamforming for replay attack detection, demonstrating its effectiveness across various devices and environments.

Findings

Achieved competitive performance on the ReMASC dataset.

Acoustic maps provide a compact, interpretable feature space.

Effective across different devices and acoustic conditions.

Abstract

Replay attacks remain a critical vulnerability for automatic speaker verification systems, particularly in real-time voice assistant applications. In this work, we propose acoustic maps as a novel spatial feature representation for replay speech detection from multi-channel recordings. Derived from classical beamforming over discrete azimuth and elevation grids, acoustic maps encode directional energy distributions that reflect physical differences between human speech radiation and loudspeaker-based replay. A lightweight convolutional neural network is designed to operate on this representation, achieving competitive performance on the ReMASC dataset with approximately 6k trainable parameters. Experimental results show that acoustic maps provide a compact and physically interpretable feature space for replay attack detection across different devices and acoustic environments.