Skip to main content
. 2021 Nov 10;11:763527. doi: 10.3389/fonc.2021.763527

Figure 2.

Figure 2

Main process of constructing a compound framework by combining pathological images with genomic data or clinical information. Convolutional neural networks are commonly used to extract image features, and then genomic features or clinical information are integrated into the full connection layer. Support vector machine (a), logistic regression (b), convolutional neural network (c) or random forest (d) can be used to establish the final multimodal fusion model.