Medicine discrimination of NIRS based on a regularized collaborative representation identification with the gabor optimizer

Abstract

Counterfeit medicines, which adversely affect patients’ health and lives, are widespread. Therefore, discriminating between genuine and counterfeit medication is becoming a critical concern. Near-infrared spectroscopy (NIRS) is a popular and effective technique that has nondestructive characteristics and is used for medicine identification. To discriminate between genuine and counterfeit medicines, a sparse signal representation model is established in the presence of spectrum crossover and overlapping. However, the sparsity of nonzero representation coefficients is low when solving L2 -norm in the representation model. To overcome this problem, a novel identification model, called regularized collaborative representation identification with the Gabor optimizer (RCRCG), is proposed in this paper. A Gabor filter is adopted to control L2 -norm for the more relevant factor vectors. Then, Lasso regulation on local identification is used to improve the accuracy of medicine discrimination. The results of experiments using NIRS samples from three datasets (with the use of erythromycin ethylsuccinate and domperidone as active substances) show that the proposed method is more effective and robust than the existing methods and its speed is twice that of the Sparse Representation-based Classification (SRC) and Class L1 -optimizer classifier with the closeness rule (C_CL1C).