Plasmonic artificial inspector for herbal medicines via surface-enhanced Raman spectroscopy and deep learning.

Area of Science:

• Analytical Chemistry
• Spectroscopy
• Artificial Intelligence

Background:

• Organoleptic examination is crucial for herbal medicine (HM) safety but is labor-intensive.
• There is a need for efficient and reliable methods to complement traditional HM inspection processes.

Purpose of the Study:

• To develop a plasmonic artificial inspector integrating surface-enhanced Raman spectroscopy (SERS) and deep learning (DL).
• To assess the accuracy and reliability of the SERS-DL system for differentiating HM species.

Main Methods:

• Utilized SERS to obtain spectral fingerprint information from HM specimens, reflecting bioactive compounds.
• Applied deep learning (DL) algorithms to analyze SERS spectra for HM identification.
• Evaluated the system's performance in differentiating 35 HM species, including those with similar appearances.

Main Results:

• The SERS-DL system achieved approximately 95% accuracy in differentiating HM species.
• The rapid data acquisition (seconds) makes SERS-DL suitable for complementary inspection.
• Demonstrated the potential for labor-saving differentiation of complex HM samples.

Conclusions:

• The synergistic integration of SERS and DL provides an accurate and efficient method for HM inspection.
• This SERS-DL approach can aid traditional organoleptic examinations and enhance HM databases.
• The technology offers a promising solution for improving the safety monitoring of herbal medicines.