A novel deep learning-based chemical image identification method of infrared spectroscopy using external perturbation

Perturbation-induced infrared spectroscopy combining deep learning-based chemical image identification method is proven to enable rapid and non-destructive identification of samples of different classes with a quite similar composition and morphological complexity with experimental validation. This study is significant for implementing advanced technology in the field of deep learning to solve chemical problems.