A Technical Review of Convolutional Neural Network-Based Mammographic Breast Cancer Diagnosis

Figure 5

The flow chart and an example of transferred MBCD models. The flow chart emphasizes transfer learning (dashed arrows) and fine-tuning, and the example comes from [64] which makes use of pretrained VGG16 for malignancy prediction. It should be noted that parameters of pretrained models are well-determined in the source domain, while fine-tuning attempts to use medical instances for further optimization of these parameters toward the target task.