Versatile Cardiovascular Signal Generation with a Unified Diffusion Transformer

2025-05-28

AI Summary

Real-time multimodal cardiovascular monitoring (PPG, ECG, BP) is hindered by wearable sensor noise and the burden of invasive acquisition. To address this, we propose UniCardio—the first unified diffusion Transformer tailored for cardiovascular signals—featuring a novel multimodal architecture that supports dynamic modality composition, integrated with continual learning and cross-modal physiological alignment. UniCardio jointly enables low-quality signal denoising, missing-modality imputation, and cross-modal synthesis (e.g., BP generation from PPG), balancing generation fidelity, out-of-distribution generalization, and clinical interpretability. Experiments demonstrate significant improvements over single-task baselines across denoising, imputation, and modality translation. Synthesized signals achieve performance on par with ground-truth measurements in abnormality detection and vital sign estimation, while maintaining robustness across unseen devices and demographic domains.

Technology Category

• Computer Vision: Multi-modal Vision
• Machine Learning: Multimodal Learning
• Intelligent Robots: Multimodal Perception & Sensor Fusion

Application Category

• User Modeling, Personalization and Recommendation: On-Device user modeling, personalization, and recommendation
• Search and Retrieval-Augmented AI: Retrieval-Augmented Generation (RAG) and multi-modal RAG
• Semantics and Knowledge: Methods to enhance, augment, integrate or synergize semantic models such as knowledge graphs and LLMs

Abstract

Cardiovascular signals such as photoplethysmography (PPG), electrocardiography (ECG), and blood pressure (BP) are inherently correlated and complementary, together reflecting the health of cardiovascular system. However, their joint utilization in real-time monitoring is severely limited by diverse acquisition challenges from noisy wearable recordings to burdened invasive procedures. Here we propose UniCardio, a multi-modal diffusion transformer that reconstructs low-quality signals and synthesizes unrecorded signals in a unified generative framework. Its key innovations include a specialized model architecture to manage the signal modalities involved in generation tasks and a continual learning paradigm to incorporate varying modality combinations. By exploiting the complementary nature of cardiovascular signals, UniCardio clearly outperforms recent task-specific baselines in signal denoising, imputation, and translation. The generated signals match the performance of ground-truth signals in detecting abnormal health conditions and estimating vital signs, even in unseen domains, while ensuring interpretability for human experts. These advantages position UniCardio as a promising avenue for advancing AI-assisted healthcare.

Problem

• Reconstructs low-quality cardiovascular signals from noisy wearables
• Synthesizes unrecorded signals using a unified generative framework
• Improves signal denoising, imputation, and translation across modalities

Innovation

• Unified diffusion transformer for multi-modal signal generation
• Specialized model architecture for diverse signal modalities
• Continual learning paradigm for varying modality combinations