Mind the Gap: Detecting Cluster Exits for Robust Local Density-Based Score Normalization in Anomalous Sound Detection
Authors: Kevin Wilkinghoff, Gordon Wichern, Jonathan Le Roux, Zheng-Hua Tan
Organizations: Department of Electronic Systems, Aalborg University, Aalborg, Denmark · Pioneer Centre for Artificial Intelligence, Copenhagen, Denmark · Mitsubishi Electric Research Laboratories (MERL), Cambridge, MA, USA
Local density-based score normalization is an effective component of distance-based embedding methods for anomalous sound detection, particularly when data densities vary across conditions or domains. In practice, however, performance depends strongly on neighborhood size. Increasing it can degrade detection accuracy when neighborhood expansion crosses cluster boundaries, violating the locality assumption of local density estimation. This observation motivates adapting the neighborhood size based on locality preservation rather than fixing it in advance. We realize this by proposing cluster exit detection, a lightweight mechanism that identifies distance discontinuities and selects neighborhood sizes accordingly. Experiments across multiple embedding models and datasets show improved robustness to neighborhood-size selection and consistent performance gains.
Figures & tables
Figure 1: Average sorted distances (left) and distance ratios (right) for BEATs embeddings of the “ToyCar” machine on the DCASE2025 dataset in the source and target domains. Pronounced distance jumps and low ratios mark cluster exits, which occur earlier in the target domain and reveal violations of locality under fixed neighborhood sizes.
Figure 2: Relative performance with respect to not applying ldn , shown as a function of the neighborhood size K . Values correspond to the ratio between the performance obtained with the respective normalization method and that obtained without ldn . Results are geometric means across all five evaluated datasets. For Direct- act , results are averaged over ten independent trials.
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.
Phurich Saengthong, Takahiro Shinozaki
Department of Information and Communications Engineering, Institute of Science Tokyo, Japan
Training-free anomalous sound detection (ASD) based on pre-trained audio embedding models has recently garnered significant attention, as it enables the detection of anomalous sounds using only normal reference data without task-specific model training or fine-tuning. However, existing embedding-based approaches almost exclusively rely on temporal mean pooling, leaving temporal pooling in training-free ASD largely unexplored. In this paper, we present the first systematic evaluation of temporal pooling strategies for training-free ASD with pre-trained audio embeddings. We propose relative deviation pooling (RDP), an adaptive pooling method that assigns larger weights to embeddings with stronger temporal deviations, investigate feature-wise non-linear aggregation using generalized mean (GeM) pooling, and examine a hybrid combination of both strategies. Experiments on five benchmark datasets demonstrate that the proposed pooling strategies consistently outperform mean pooling and achieve state-of-the-art performance for training-free ASD, including results that surpass previously reported trained systems and ensembles on the DCASE2025 ASD dataset.
Kevin Wilkinghoff, Sarthak Yadav, Zheng-Hua Tan
Department of Electronic Systems, Aalborg University, Aalborg, Denmark · Pioneer Centre for Artificial Intelligence, Denmark
First-shot anomalous sound detection in DCASE Challenge Task 2 must flag anomalies of unseen machine types with a single threshold, without knowing whether a test clip comes from the data-rich source domain (990 normal training clips) or the data-scarce target domain (10). Two organizer-reported problems remain open: source- and target-domain AUC are negatively correlated across systems, and development-set performance does not predict evaluation-set performance. We address both with a training-free post-hoc layer over frozen audio embeddings: (i) per-domain quantile calibration shrunk toward a pooled map by a prior strength m, tracing a source/target balance frontier, and (ii) a label-free cross-validated domain-balance criterion that ranks candidate configurations from training normals only, paired with a coarse development-labeled viability veto. On DCASE 2025, the criterion rank-predicts the official evaluation score across a 45-configuration grid (Spearman rho = +0.91; family-block bootstrap 95% CI [+0.83, +0.95]) while development score is uninformative (+0.06). Criterion-based selection raises the evaluation score from 55.83 to 59.34 (jackknife CI [2.2, 4.8]) and, on an extended grid, to 61.05 -- retrospectively fourth of 35 teams. Replicating on DCASE 2023 and 2024 bounds the claim: development score is uninformative in all three years and degenerate configurations recur (vetoed every time), but under family-clustered uncertainty the criterion's predictive evidence survives only in 2025; in both replication years a fixed full-equalization default matches or beats criterion-based selection. A DCASE 2026 forward test is frozen before the 2026 evaluation ground truth is released; all headline numbers are reproduced by the official evaluator.