AI-Driven Cardiorespiratory Signal Processing: Separation, Clustering, and Anomaly Detection

TL;DR

This work addresses the challenge of effective AI-driven analysis of cardiorespiratory signals by introducing a dedicated heart–lung sound dataset (HLS-CMDS) and a suite of advanced algorithms. It combines generative AI for blind source separation (LingoNMF, XVAE-WMT), a physics-inspired clustering method (Chem-NMF), and a quantum-classical detector (QuPCG) to separate, cluster, and identify anomalies in heart and lung sounds. The study also surveys sensing technologies from MEMS to integrated quantum photonics, arguing for a future where AI and next-generation sensors enable intelligent, chip-scale diagnostics. Key findings show measurable gains in separation metrics, interpretable latent representations, robust clustering, and high abnormal-pattern classification accuracy, highlighting the potential of AI-augmented biosensing for healthcare diagnostics.

Abstract

This research applies artificial intelligence (AI) to separate, cluster, and analyze cardiorespiratory sounds. We recorded a new dataset (HLS-CMDS) and developed several AI models, including generative AI methods based on large language models (LLMs) for guided separation, explainable AI (XAI) techniques to interpret latent representations, variational autoencoders (VAEs) for waveform separation, a chemistry-inspired non-negative matrix factorization (NMF) algorithm for clustering, and a quantum convolutional neural network (QCNN) designed to detect abnormal physiological patterns. The performance of these AI models depends on the quality of the recorded signals. Therefore, this thesis also reviews the biosensing technologies used to capture biomedical data. It summarizes developments in microelectromechanical systems (MEMS) acoustic sensors and quantum biosensors, such as quantum dots and nitrogen-vacancy centers. It further outlines the transition from electronic integrated circuits (EICs) to photonic integrated circuits (PICs) and early progress toward integrated quantum photonics (IQP) for chip-based biosensing. Together, these studies show how AI and next-generation sensors can support more intelligent diagnostic systems for future healthcare.

Figures (29)

• Figure 1: Biosensor development timeline. It highlights key technological milestones from the first blood oxygen biosensor, integrated circuits, and quantum sensors, to recent advances toward integrated quantum–photonic systems and AI-assisted lab-on-chip biosensing.
• Figure 1: (a) MEMS operating principle C71ref90; (b) Cross-sectional structure of a MEMS microphone C71ref90; (c) Typical MEMS microphone assembly C71ref70; (d) Digital MEMS microphone application schematic C71ref70; (e) Commercial MEMS microphone packages C71ref88.
• Figure 1: Diagram of the data acquisition process.
• Figure 1: PL-NMF overview: (a) Proposed algorithm; (b) Internal structure of the “Heart Sound NMF” and “Lung Sound NMF” blocks.
• Figure 1: Energy profile of a reaction progress: (a) Reactants, intermediate, and products. The two transition states correspond to the energy maxima, with activation free energies $\Delta G^{\ddagger}_{1}$ and $\Delta G^{\ddagger}_{2}$ indicated by vertical arrows. The overall free energy change $\Delta G$ is shown between reactants and products; (b) The effect of catalyst on lowering the activation barrier.

Theorems & Definitions (21)

• Theorem 5.4.1
• Definition 5.4.1: Auxiliary Function
• Lemma 5.4.1
• Theorem 5.4.2
• Definition 5.4.2: Energy Barrier
• Definition 5.4.3: Boltzmann Probability
• Lemma 5.4.2
• Theorem 5.4.3
• Lemma 5.4.3
• Theorem 5.4.4