Fig. 1
From: Convolutional neural network models for cancer type prediction based on gene expression
Illustration of three CNN models. a 1D-CNN with input as a vector format with 7100 genes. b 2D-Vanilla-CNN, with an input reformatted as a 100 × 71 matrix, and one convolution layer. c 2D-Hybrid-CNN, similar input as in (b) but with two parallel convolution layers, vertical and horizontal, as in (a)