Search papers, labs, and topics across Lattice.
2
0
3
This work frames this as Domain-Incremental Learning over acoustic environments and presents the first continual learning benchmark for multi-channel replay speech detection, evaluating a state-of-the-art beamformer-based detector over all 24 environment orderings of the ReMASC corpus with five seeds.
Acoustic maps offer a compact and physically interpretable feature space that allows lightweight CNNs to effectively detect replay attacks, even across diverse devices and acoustic environments.