Deciphering quality decline in Peucedani Radix: An interpretable deep learning framework integrating multimodal analysis across the supply chain

Key Points

• The study aims to assess the quality of Peucedani Radix and identify factors leading to its decline.
• Conducted comprehensive quality assessment across the supply chain.
• Utilized chemical, molecular, and spectroscopic analyses.
• Developed an interpretable deep learning model for rapid authentication.
• Achieved over 90% accuracy in quality assessment using the deep learning model.
• Identified multiple factors contributing to quality decline across the supply chain.
• Proposed a new quality control paradigm combining verification and on-site rapid screening.