Fluorescence background removal method for biological Raman spectroscopy based on empirical mode decomposition

Area of Science:

• Biomedical Optics
• Signal Processing
• Spectroscopy

Background:

• Fluorescence background in biological Raman spectroscopy generates inaccurate intensity data.
• Existing baseline correction methods often require parameter selection, limiting their adaptability.

Purpose of the Study:

• To evaluate the Empirical Mode Decomposition (EMD) method for fluorescence background removal in biological Raman spectra.
• To assess EMD's performance against established algorithms like the Vancouver Raman Algorithm (VRA).

Main Methods:

• Application of Empirical Mode Decomposition (EMD) for adaptive signal processing and baseline correction.
• Validation using synthetic Raman spectra with varying signal-to-noise ratios (SNR).
• Comparative analysis with the Vancouver Raman Algorithm (VRA) using biological skin Raman spectra.

Main Results:

• EMD achieved a correlation coefficient greater than 0.92 for synthetic Raman spectra.
• Comparison with VRA on skin spectra yielded a low Mean Square Error (MSE) of 0.001554.
• EMD demonstrated robust performance without requiring parameter selection.

Conclusions:

• Empirical Mode Decomposition (EMD) is a highly effective and parameter-free method for correcting fluorescence background in biological Raman spectra.
• EMD offers a promising alternative for enhancing the quality of Raman spectral data in biomedical research.
• The adaptive nature of EMD makes it suitable for analyzing complex, non-stationary biological signals.