RepAugment: Input-Agnostic Representation-Level Augmentation for Respiratory Sound Classification

TL;DR

This paper introduces RepAugment, a novel data augmentation method for respiratory sound classification that improves accuracy, especially for minority disease classes, by bridging the gap between speech and lung sound representations.

Contribution

The paper proposes RepAugment, a representation-level augmentation technique compatible with waveform pretrained models, outperforming traditional spectrogram-based methods like SpecAugment.

Findings

RepAugment outperforms SpecAugment in accuracy.

Significant improvement in minority disease class detection, up to 7.14%.

Demonstrates the effectiveness of input-agnostic augmentation for respiratory sounds.

Abstract

Recent advancements in AI have democratized its deployment as a healthcare assistant. While pretrained models from large-scale visual and audio datasets have demonstrably generalized to this task, surprisingly, no studies have explored pretrained speech models, which, as human-originated sounds, intuitively would share closer resemblance to lung sounds. This paper explores the efficacy of pretrained speech models for respiratory sound classification. We find that there is a characterization gap between speech and lung sound samples, and to bridge this gap, data augmentation is essential. However, the most widely used augmentation technique for audio and speech, SpecAugment, requires 2-dimensional spectrogram format and cannot be applied to models pretrained on speech waveforms. To address this, we propose RepAugment, an input-agnostic representation-level augmentation technique that outperforms SpecAugment, but is also suitable for respiratory sound classification with waveform pretrained models. Experimental results show that our approach outperforms the SpecAugment, demonstrating a substantial improvement in the accuracy of minority disease classes, reaching up to 7.14%.