An Enhanced Audio Feature Tailored for Anomalous Sound Detection Based on Pre-trained Models

2025-08-21

AI Summary

To address the challenges of uncertain anomaly localization and redundant audio noise in machine abnormal sound detection (ASD), this paper proposes a parameter-free uniform filter bank–based feature enhancement method. First, a full-spectrum, equally spaced filter bank is designed to achieve balanced spectral representation. Second, generic acoustic features are extracted using a pre-trained model, and a learnable-parameter-free feature enhancement mechanism is introduced to facilitate efficient knowledge transfer to the ASD task. The method jointly achieves noise suppression and discriminative modeling without requiring task-specific parameter tuning. Evaluated on the DCASE 2024 Challenge dataset, it significantly improves detection accuracy and robustness to acoustic noise. This work establishes a new paradigm for lightweight, transferable ASD systems, demonstrating strong generalization across diverse industrial machinery scenarios while maintaining computational efficiency.

Technology Category

• Machine Learning: Transfer, Domain Adaptation, Multi-Task Learning
• Data Mining & Knowledge Management: Anomaly/Outlier Detection
• Natural Language Processing: Learning & Optimization for NLP

Application Category

• Responsible Web: Machine-in-the-loop, human agency and autonomy
• User Modeling, Personalization and Recommendation: On-Device user modeling, personalization, and recommendation
• Systems and Infrastructure for Web, Mobile and WoT: Applied ML and AI for Web-based mobile applications

Abstract

Anomalous Sound Detection (ASD) aims at identifying anomalous sounds from machines and has gained extensive research interests from both academia and industry. However, the uncertainty of anomaly location and much redundant information such as noise in machine sounds hinder the improvement of ASD system performance. This paper proposes a novel audio feature of filter banks with evenly distributed intervals, ensuring equal attention to all frequency ranges in the audio, which enhances the detection of anomalies in machine sounds. Moreover, based on pre-trained models, this paper presents a parameter-free feature enhancement approach to remove redundant information in machine audio. It is believed that this parameter-free strategy facilitates the effective transfer of universal knowledge from pre-trained tasks to the ASD task during model fine-tuning. Evaluation results on the Detection and Classification of Acoustic Scenes and Events (DCASE) 2024 Challenge dataset demonstrate significant improvements in ASD performance with our proposed methods.

Problem

• Detecting anomalous sounds in machines with uncertain anomaly locations
• Removing redundant information and noise from machine audio data
• Improving anomaly detection performance across all frequency ranges

Innovation

• Evenly distributed filter banks feature
• Parameter-free feature enhancement approach
• Pre-trained model knowledge transfer strategy