A rapid method for distinguishing similar gelatins based on terahertz spectrum

Distinguishing similar gelatins are essential for pharmaceutical administration and therapy based on traditional Chinese medicine. Differed from the conventional studies, this study focuses on a reaction-free terahertz approach to identify gelatins made of asinine, bovine, and porcine skins. Samples in tablets form are measured by terahertz time-domain spectroscopy and the calculated optical parameters are processed by a principal component analysis algorithm. Additionally, linear support vector machines are used to classify gelatin varieties and the performance is discussed with respect to the number of principal components kept for modelling. It is founded that three principal components are optimal to identify similar gelatins in a reduced feature space and the extinction coefficient is more efficient in distinguishing three different gelatins. This study proves the feasibility of THz time-domain spectroscopy in gelatin administration in traditional Chinese medicine.