This is an intel-extended caffe based 3D faster RCNN RPN training framework, which we believe is the first training framework that makes 3D faster RCNN RPN with 150-layer Deep Convolutional Network converged in CT images.
The model has achieved good performance on Alibaba TIANCHI Healthcare AI Competition data (medical imaging prediction of lung nodule). You are welcome to modify it to GPU version.
This open-source project is developed by Shenzhen Yiyuan Intelligence Tech Co., LTD and Hong Kong Baptist University (HKBU) GPU High Performance Computing Laboratory.
The 3D RPN network : models/tianchi/VGG16/faster_rcnn_end2end/train.prototxt
Input data layer: lib/roi_data_layer/layer.py
The training data stored in directory :data/tianchi/data
[[ 201. , 242. , 112. , 8.12129222],
[ 231. , 390. , 132. , 4.43397444]]
...
The first three of each line are z, x, y coordinates, and the fourth number is the nodule size(mm)
python -u train_net.py --solver ../models/tianchi/VGG16/faster_rcnn_end2end/solver.prototxt --imdb_train tianchi_train --imdb_val tianchi_val --iters 70000 --cfg ../experiments/cfgs/faster_rcnn_end2end.yml --rand
the model output in directory :output/faster_rcnn_end2end/tianchi_train
python -u val_net.py --solver ../models/tianchi/VGG16/faster_rcnn_end2end/solver_val.prototxt --imdb_train tianchi_train --imdb_val tianchi_val --iters 70000 --cfg ../experiments/cfgs/faster_rcnn_end2end.yml --rand
python -u test_net.py --def ../models/tianchi/VGG16/faster_rcnn_end2end/test.prototxt --net ../output/faster_rcnn_end2end/tianchi_faster_rcnn_iter_2204.caffemodel --imdb tianchi_test --cfg ../experiments/cfgs/faster_rcnn_end2end.yml --max_per_image 1
Yu Wu : YiYuan Intelligent co-founder
Shaohuai Shi : Hong Kong Baptist University, Phd
Xiaochen Chen : Hong Kong University of Science and Technology, Master