Table 1 Dataset parameters and CNN classification speeds. Column definitions: Acquisition, motorized stage or freehand motion used for spatial scanning; Grating, lines/mm; n, number of technical replicates performed, except for tissue biopsies where it is the number of biopsy samples imaged; x, number of spatial steps resulting in 2D spatial-spectral images employed for training; y, spatial size of the measured line on the spectrometer camera in pixels (each of dimension 16 μm × 16 μm); λ, spectral dispersion on the camera in pixels; CNN speed, potential classification speed calculated from the time taken for analysis when the CNN is applied to the dataset.

From: Deep learning applied to hyperspectral endoscopy for online spectral classification