Sub-Band Spectral Matching with Localized Score Aggregation for Robust Anomalous Sound Detection

TL;DR

This paper introduces BEAM, a novel sub-band spectral matching method with localized score aggregation that enhances robustness and discriminability in unsupervised anomalous sound detection, especially in noisy environments.

Contribution

BEAM employs sub-band neighbor retrieval and adaptive score fusion to reduce normal-score variance and improve detection accuracy without task-specific training.

Findings

Outperforms existing methods on DCASE benchmarks

Robust to noise and domain shifts

Enhances detection without task-specific training

Abstract

Detecting subtle deviations in noisy acoustic environments is central to anomalous sound detection (ASD). A common training-free ASD pipeline temporally pools frame-level representations into a band-preserving feature vector and scores anomalies using a single nearest-neighbor match. However, this global matching can inflate normal-score variance through two effects. First, when normal sounds exhibit band-wise variability, a single global neighbor forces all bands to share the same reference, increasing band-level mismatch. Second, cosine-based matching is energy-coupled, allowing a few high-energy bands to dominate score computation under normal energy fluctuations and further increase variance. We propose BEAM, which stores temporally pooled sub-band vectors in a memory bank, retrieves neighbors per sub-band, and uniformly aggregates scores to reduce normal-score variability and improve discriminability. We further introduce a parameter-free adaptive fusion to better handle diverse temporal dynamics in sub-band responses. Experiments on multiple DCASE Task 2 benchmarks show strong performance without task-specific training, robustness to noise and domain shifts, and complementary gains when combined with encoder fine-tuning.