Smart Cardiovascular Diagnosis Using Optical Biosignals and Capsule Intelligence Networks

Abstract

Cardiovascular diseases (CVDs) remain among the most significant causes of global mortality, necessitating the development of intelligent diagnostic systems capable of providing accurate and timely detection of cardiac abnormalities. Optical bio signals, particularly photoplethysmography (PPG) and related non-invasive sensing modalities, have emerged as promising alternatives to traditional diagnostic approaches due to their ease of acquisition, cost-effectiveness, and suitability for continuous health monitoring. However, the complex physiological patterns embedded within optical bio signals present substantial challenges for conventional analytical methods, limiting diagnostic reliability and predictive performance. To address these challenges, this study proposes a Smart Cardiovascular Diagnosis Framework Using Optical Biosignals and Capsule Intelligence Networks (SCD-OBCIN) for intelligent detection and classification of cardiovascular abnormalities. The proposed framework integrates advanced optical bio signal preprocessing, multilevel feature extraction, and capsule-based deep learning architectures to capture hierarchical spatial and temporal relationships within physiological signals. Capsule Intelligence Networks effectively preserve feature dependencies and improve representation learning, enabling accurate recognition of subtle cardiovascular abnormalities that may be overlooked by traditional neural models. Furthermore, adaptive learning mechanisms enhance robustness against signal noise, motion artifacts, and inter-patient variability.