Abstract: To address the issues of the continuous emergence of adulterants in the Radix Fici Hirtae market and the subjectivity and experience dependence of traditional identification methods, high-performance liquid chromatography (HPLC) fingerprinting was combined with artificial intelligence (AI) technology. Samples of Radix Fici Hirtae and its adulterants were collected, and their HPLC fingerprints were established. An improved YOLO11 algorithm was introduced into the analysis of traditional Chinese medicine fingerprints, enabling automatic and accurate recognition of characteristic chromatographic peaks and learning of their multi-dimensional morphological features. Consequently, an end-to-end intelligent identification model was established, ranging from raw chromatograms to authenticity discrimination. The results demonstrated that the constructed YOLO11 identification model was feasible, achieving an accuracy of 99.19%, recall of 100.00%, precision of 98.25%, specificity of 98.53%, and an F1 score of 99.12%. These metrics indicated that the model achieved an optimal balance between precision and comprehensiveness, and could accurately distinguish Radix Fici Hirtae from non-Radix Fici Hirtae samples.

Key words: high-performance liquid chromatography fingerprint, Radix Fici Hirtae, artificial intelligence, YOLO11, identification