The research on artificial intelligence-based Deer Velvet Antler traceability model based on emergent data features

Area of Science:

• Food Science
• Analytical Chemistry
• Computational Biology

Background:

• Deer velvet antler (DVA) is nutritionally valuable but faces market adulteration due to complex origins.
• Accurate species identification is crucial for DVA quality and traceability.

Purpose of the Study:

• To develop an efficient and accurate method for identifying different species of DVA.
• To enhance classification accuracy and traceability in the DVA market.

Main Methods:

• Collected 120 DVA samples from four species: Sika Deer (SVA), Wapiti (WVA), Reindeer (RVA), and Moose (MVA).
• Extracted multidimensional features (color, texture, odor, composition) using Computer Vision, Electronic Nose, and HPLC.
• Developed a Whale Optimization Algorithm-Random Forest (WOA-RF) classification model based on 162 discriminative factors.

Main Results:

• Identified 162 key discriminative factors through multivariate statistical analysis.
• The WOA-RF model achieved a 100% success rate in classifying DVA species.
• Demonstrated emergent effects from multidimensional feature fusion and optimization algorithms.

Conclusions:

• The proposed intelligent algorithm offers highly efficient DVA species identification.
• This approach surpasses single-technique limitations, providing technical support for food species identification.
• Enhances traceability and quality control for valuable DVA products.