Image-Based Quantitative Root Phenotyping Enables Reliable Origin Classification of Wild-Simulated Ginseng

Reliable origin authentication is needed for wild-simulated ginseng (WSG; Panax ginseng C. A. Meyer) because lower-priced imported products may be mislabeled as domestically produced Korean WSG. This study evaluated whether image-derived root morphology can provide an objective framework for distinguishing Korean WSG from Chinese-origin ginseng samples. A total of 248 specimens (176 Korean and 72 Chinese-origin samples) were imaged, segmented into root crown and whole-root regions, and quantified using 41 morphological traits, comprising 12 crown traits and 29 whole-root traits. Trait differences and discriminative patterns were assessed using FDR-controlled statistical testing, multivariate analysis, and classification modeling. After FDR correction, 35 of 41 traits remained robustly different between origins, including 9 of 12 crown traits and 26 of 29 whole-root traits. Chinese-origin samples generally showed larger size- and thickness-related values, particularly for area, perimeter, diameter, surface area, and volume. The FDR-selected traits supported origin-associated multivariate separation and high classification performance, with accuracy ranging from 98.8–99.6%. These findings indicate that quantitative root phenotyping can support non-destructive origin-group screening of ginseng and provide a practical basis for automated authentication and quality-control systems, although external validation across broader cultivation and environmental conditions remains necessary.