论文:end to end training for whole image breast cancer screening using an all convolutional design
数据集:CBIS-DDSM (CC+MLO)*(right +left) one person ==》 (753钙化)+(891肿块) 数据增扩
patch set
- s1: 1 ROI +1背景 in the same image
- s10:ROI overlapping raito of 0.9, 10 patch and 10 background in the same image
任务:patch训练 (5类:背景,钙化良性,钙化恶性,肿块良性,肿块恶性)
二分类:正常(0-2),异常(4-6) (3不参与分类)
定位:CAM
网络结构: Vggnet /Resnet+two classification
一、训练流程:
中心思想:patch 训练分类器,再用whole image 去(训练+测试)