Search

Medicinal plants have a variety of values and are an important source of new drugs and their lead compounds. They have played an important role in the treatment of cancer, AIDS, COVID-19 and other major and unconquered diseases. However, there are problems such as uneven quality and adulteration. Therefore, it is of great significance to find comprehensive, efficient and modern technology for its identification and evaluation to ensure quality and efficacy. In this study, deep learning, which is superior to conventional identification techniques, was extended to the identification of the part and region of the medicinal plant Paris polyphylla var. yunnanensis from the perspective of spectroscopy. Two pattern recognition models, partial least squares discriminant analysis (PLS-DA) and support vector machine (SVM), were established, and the overall discrimination performance of the three types of models was compared. In addition, we also compared the effects of different sample sizes on the discriminant performance of the models for the first time to explore whether the three models had sample size dependence. The results showed that the deep learning model had absolute superiority in the identification of medicinal plant. It was almost unaffected by factors such as data type and sample size. The overall identification ability was significantly better than the PLS-DA and SVM models. This study verified the superiority of the deep learning from examples, and provided a practical reference for related research on other medicinal plants. [ABSTRACT FROM AUTHOR]

[Display omitted] • Resnet with 2DCOS images for identifying growth years was first proposed. • The proposed Resnet model can realize identification with 100% accuracy. • A total of 4725 2DCOS images were produced for identification analysis. • Synchronous 2DCOS has better identification performance in complex systems. Paris polyphylla var. yunnanensis , as perennial plants, its quality is closely related to growth period. Different harvest years determine the dry matter accumulation of its medicinal parts and the dynamic accumulation of active ingredients, as well as its economic value and medicinal value. Therefore, it is necessary to establish a systematic evaluation method for the identification and evaluation of P. polyphylla var. yunnanensis with different growth years. Deep learning has a powerful ability in recognition. This study extends it to the identification analysis of medicinal plants from the perspective of spectrum. For the first time, two-dimensional correlation spectroscopy (2DCOS) based on the attenuated total reflection Fourier transformed infrared spectroscopy (ATR-FTIR) combined with residual neural network (Resnet) was used to identify growth years. 525 samples were collected, 4725 2DCOS images were drawn, and the dry matter accumulation in rhizomes of different growth years and different sampling sites were briefly analyzed. The results show that the eight-year-old P. polyphylla var. yunnanensis in Dali has higher economic value and medicinal value. The synchronous 2DCOS models based on ATR-FTIR can realize the identification of growth years with accuracy of 100%. Synchronous 2DCOS are more suitable for the identification of medicinal plants with complex systems. 2DCOS images with different colors and second derivative processing cannot optimize the modeling results. In summary, the method we established is innovative and feasible. It not only solved the identification of growth years, expanded the application field of deep learning, but could also be extended to further research on other medicinal plants. [ABSTRACT FROM AUTHOR]

• 2DCOS with ResNet for the identification of Paris species was first proposed. • Multiple feature bands fusion strategy of 2DCOS was first proposed for modeling. • A total of 5040 2DCOS images were produced for identification analysis. • Synchronous 2DCOS has better identification performance in complex systems. Spectral analysis has the characteristics of fast and nondestructive. In order to conform to the development of the times, a practical method beyond the traditional spectral analysis was established. For the first time, the two-dimensional correlation spectroscopy (2DCOS) images of Fourier-transform mid-infrared spectroscopy combined with the Residual Neural Network (ResNet) was used for the identification and analysis of 12 Paris species, and the second derivative 2DCOS rarely involved in previous researchers was established. Besides, the fusion strategy of 2DCOS images based on feature bands was first proposed for modeling analysis. From the results, (1) 2DCOS combined with ResNet can successfully identify 12 Paris species. (2) 2DCOS is a powerful tool for identification, whether it is used for image visual analysis or modeling analysis. (3) Compared with asynchronous 2DCOS, synchronous 2DCOS is more suitable for the identification and analysis of complex mixed systems such as traditional Chinese medicine. (4) The modeling based on feature bands fusion strategy of 2DCOS has better model performance and is also suitable for the analysis of small samples. To sum up, what we proposed is an innovative and feasible method with wide applicability, which can not only solve the problem of identifying Paris , provide ideas and methods for the selection of spectral types and feature bands, but also provide practical reference for the research in analytical chemistry and other related fields. [ABSTRACT FROM AUTHOR]