Table 1. Type of tissue, the colour of label and number of spectra obtained for this study
After labelling, the following steps were carried out:
Application of the network to distinguish between benign and malignant lymphoid tissue types. To do so, the neural network is trained with two sets of spectral data (e.g., FL vs DLBCL, non-GCB). The evaluation of the deep classifier is done on the second set of spectra containing the respective classes.
Classification of more than two types (multiclass) of neoplastic and reactive lymphoid tissue.