Improving Anomalous Sound Detection through Pseudo-anomalous Set Selection and Pseudo-label Utilization under Unlabeled Conditions

TL;DR

This paper proposes a novel approach for anomalous sound detection that leverages pseudo-anomalous set selection and pseudo-labeling to improve detection accuracy in unlabeled and scarce data scenarios, demonstrating significant performance gains.

Contribution

It introduces an integrated pipeline combining anomaly score-based data selection, triplet learning for pseudo-labeling, and iterative training to enhance ASD under unlabeled conditions.

Findings

Achieves over 6.6 points increase in average AUC in unlabeled settings.

Improves detection accuracy by refining pseudo-labels and data selection.

Demonstrates robustness and practicality in industrial scenarios with minimal labels.

Abstract

This paper addresses performance degradation in anomalous sound detection (ASD) when neither sufficiently similar machine data nor operational state labels are available. We present an integrated pipeline that combines three complementary components derived from prior work and extends them to the unlabeled ASD setting. First, we adapt an anomaly score based selector to curate external audio data resembling the normal sounds of the target machine. Second, we utilize triplet learning to assign pseudo-labels to unlabeled data, enabling finer classification of operational sounds and detection of subtle anomalies. Third, we employ iterative training to refine both the pseudo-anomalous set selection and pseudo-label assignment, progressively improving detection accuracy. Experiments on the DCASE2022-2024 Task 2 datasets demonstrate that, in unlabeled settings, our approach achieves an average AUC increase of over 6.6 points compared to conventional methods. In labeled settings, incorporating external data from the pseudo-anomalous set further boosts performance. These results highlight the practicality and robustness of our methods in scenarios with scarce machine data and labels, facilitating ASD deployment across diverse industrial settings with minimal annotation effort.