A Chinese Heart Failure Status Speech Database with Universal and Personalised Classification

2025-08-12

AI Summary

This study addresses a critical gap in non-invasive heart failure (HF) monitoring by investigating, for the first time, whether Mandarin speech syllables inherently encode HF status information. Method: We constructed the first paired Mandarin speech database comprising pre- and post-hospitalization recordings from the same HF patients. To mitigate inter-speaker variability, we proposed a personalized “paired” classification framework and designed an adaptive frequency filter to extract discriminative spectral features. Statistical hypothesis testing and machine learning analyses were conducted to evaluate classification performance. Contribution/Results: Results demonstrate that Mandarin speech significantly reflects HF status changes, with the paired framework substantially improving robustness over conventional patient-wise classification. This work establishes the feasibility of using Mandarin speech for passive, non-invasive HF monitoring. All data and source code are publicly released to support reproducibility and further research.

Technology Category

• Natural Language Processing: Speech
• Data Mining & Knowledge Management: Conversational Systems for Recommendation & Retrieval
• Machine Learning: Multimodal Learning

Application Category

• Search and Retrieval-Augmented AI: Multilingual and cross-lingual Web search
• Web Mining and Content Analysis: Mining multimedia, multimodal, multilingual, cross-lingual Web data
• User Modeling, Personalization and Recommendation: User modeling and simulation for interactive and conversational systems

Abstract

Speech is a cost-effective and non-intrusive data source for identifying acute and chronic heart failure (HF). However, there is a lack of research on whether Chinese syllables contain HF-related information, as observed in other well-studied languages. This study presents the first Chinese speech database of HF patients, featuring paired recordings taken before and after hospitalisation. The findings confirm the effectiveness of the Chinese language in HF detection using both standard 'patient-wise' and personalised 'pair-wise' classification approaches, with the latter serving as an ideal speaker-decoupled baseline for future research. Statistical tests and classification results highlight individual differences as key contributors to inaccuracy. Additionally, an adaptive frequency filter (AFF) is proposed for frequency importance analysis. The data and demonstrations are published at https://github.com/panyue1998/Voice_HF.

Problem

• Detecting heart failure using Chinese speech analysis
• Validating Chinese syllables for HF-related information identification
• Addressing individual differences in HF speech classification accuracy

Innovation

• First Chinese HF speech database with paired recordings
• Personalised pair-wise classification for speaker-decoupled baseline
• Adaptive frequency filter for frequency importance analysis