Toward-Automated Etalon Artifacts Elimination in Dual-Wavelength Raman Spectroscopy for In Vivo Applications

Confocal Raman spectroscopy has emerged as a powerful tool for noninvasive chemical analysis of human skin, offering unparalleled molecular specificity. Despite its promise, the pervasive issue of etalon artifacts─caused by multiple reflections and interference in back-illuminated charge-coupled devices─significantly hinders its application, particularly in pigmented skin, which is prevalent in Asia. Addressing this critical challenge, we present a toward-automated digital processing workflow that combines independent component analysis and continuous wavelet transform to effectively identify and eliminate etalon artifacts. By leveraging the consistent fringe patterns generated in a dual-wavelength excitation Raman system, our method isolates the specific frequency and wavelength ranges impacted by etalon artifacts without requiring additional calibration or manual parameter tuning. Validation on in vivo Raman spectra from diverse skin samples demonstrates that this approach not only removes etalon artifacts with high precision but also preserves critical spectral features, establishing a robust foundation for accurate and detailed skin analysis.