A comprehensive survey of biomedical optical image reconstruction using physics-informed deep learning techniques and their applications: Status, challenges and perspectives

Accurate diagnosis and disease monitoring in healthcare is the key concern in biomedical imaging. Established Deep learning models have shown promise, but alone lacks in interpretability and physics grounding. This Comprehensive survey presents an extensive review on state-of-the art techniques in biomedical optical image reconstruction using Physics informed deep learning (PIDL) approaches whose focuses on applications including biomedical image reconstruction, diffuse optical tomography (DOT), and fluorescence microscopy. various physics guided machine learning and deep learning models examines data-driven models with first principle of physics to enhance image construction quality, reduce data requirements and ensure physically consistent outputs . overview of imaging modalities such as MRI, CT, and ultrasound are analyzed. We explore the impact of PIDL on image reconstruction, segmentation, enhancement, and anomaly detection, demonstrating its effectiveness in addressing noise, resolution, and diagnostic accuracy challenges. Hybrid models that combine PIDL with deep learning techniques, improved benchmark datasets, and scalable algorithms for real-time applications imply research future directions. This paper aims to serve as a valuable reference for researchers seeking to understand the current scenario and future prospective of biomedical optical image reconstruction utilizing physics guided deep learning.