Breast Cancer Research

Multi-task transfer learning DCNN for classification of malignant and benign breast masses in mammography. 

2D t-SNE maps of the training samples obtained from the single-task (STTL) and multi-task (MTTL) approaches for transfer networks at four fully connected layers. The color legend indicates the two mass classes in the digitized screen-film (SFM, DDSM) and digital (DM) data sets. 

Overview of the CNN structures used in the multi-stage transfer learning. (a) ImageNet trained CNN with five convolutional layers and four fully connected layers. (b) Stage 1 transfer learning using mammography data. Two fully connected layers (F4 and F5) are added to the ImageNet structure in (a). (c) Stage 2 transfer learning using DBT data. Note that (b) and (c) show three strategies of fine-tuning by freezing the CNN at different layers.