Fig. 1: Schematic illustration of deep learning-based Raman spectra analysis for CNT identification.
From: High-speed identification of suspended carbon nanotubes using Raman spectroscopy and deep learning
a Implementation of high-speed Raman imaging on a fork-like growth substrate. b Generation of unlabeled Raman spectra. c Large labeled datasets organized into three classes: S-CNTs, M-CNTs, and empty. d Deep learning model. e Classification of individual spectra using the model. f CNT identification.
