Fritillaria spp. has an extremely high edible and medicinal value. Different parts of it exhibit significant variations in medicinal efficacy. To rapidly and accurately identify the origin and adulteration of Fritillaria spp., a handheld near-infrared spectrometer was combined with a convolutional neural network (CNN) to establish an efficient and convenient quality assessment method. First, for the origin of Fritillaria spp., the CNN could achieve high accuracy, with 100 ± 0%. The features contributing to the origin of Fritillaria spp. were visualized using gradient-weighted class activation mapping (Grad-CAM). For the adulteration of Fritillaria spp., compared with partial least squares regression (PLSR), the CNN yielded the best performance, with the R² of the test set being 0.9897. Additionally, to improve the interpretability of the adulteration model, a CNN model was established using data whose dimensions had been reduced by PCA (PCA–CNN), which also achieved an R² of 0.9876. The features extracted by PCA focused on 1400–1500 nm, which was consistent with Grad-CAM. The visualization of Grad-CAM and the adulteration detection model achieved mutual validation, showing the effectiveness of both methods in analyzing the samples. The experimental results demonstrated that the integration of a handheld near-infrared spectrometer with a CNN enabled both reliable authentication of Fritillaria spp. geographical origins and quantitative determination of adulteration levels, establishing a novel analytical framework for rapid quality evaluation of Fritillaria spp. materials.

The seeds of Cichorium glandulosum Boiss. et Huet. (CS) are known for various effects. However, the research on the establishment of quality evaluation standards for CS is currently limited. Therefore, this study employed Ultra Performance Liquid Chromatography–Tandem Mass Spectrometry (UPLC-MS/MS) to analyze the components of CS. Forty-nine compounds were identified through manual analysis and database comparison. The components were then verified using HPLC and standards. Additionally, 19 batches were collected to establish the fingerprint chromatogram. Five major chemical components were selected for subsequent analysis. MIR, combined with three variable selection algorithms and three preprocessing methods, was used to build prediction models. For the three indexes of Chlorogenic Acid, 1,4-Dicaffeoylquinic Acid, and 1,5-Dicaffeoylquinic Acid, the R² values for both the training set and test set were above 0.9, the RPD values were all greater than 2.5, and the RER values were greater than 10. This indicated that the combination of mid-infrared spectroscopy and chemometrics had excellent model applicability and prediction performance for these three indexes. A quality evaluation system has been initially established, laying a foundation for research on quality evaluation of CS.