Review of Coherent Anti-Stokes Raman Scattering Nonresonant Background Removal and Phase Retrieval Approaches: From Experimental Methods to Deep Learning Algorithms

Abstract

Coherent anti-Stokes Raman spectroscopy (CARS) is a nonlinear optical technique widely utilized for vibrational imaging and molecular characterization in fields such as chemistry, biology, medicine, and materials science. Despite the high signal intensity provided by CARS, the nonresonant background (NRB) can obscure valuable molecular fingerprint information. Therefore, effective NRB removal and phase retrieval are essential for achieving precise spectral analysis and accurate material characterization. This review provides a comprehensive overview of the evolution of CARS-NRB removal and phase retrieval methods, tracing the transition from classical experimental techniques and numerical algorithms to cutting-edge deep learning models. The discussion evaluates the strengths and limitations of each approach and explores future directions for integrating deep learning to improve phase retrieval accuracy and NRB removal efficiency in CARS applications.