This code extends py-faster-rcnn by adding:
- ResNet implementation.
- Online Hard Example Mining.
- Caltech Dataset train and test interface.
- Add RPN tools.
- The faster rcnn code is based on py-faster-rcnn.
- The ohem code is based on ohem.
- To reduce the memory usage, we use batchnorm layer in Microsoft's caffe
- The caltech dataset interface based on GBJim-py-faster-rcnn
The installation and useage are same as Faster R-CNN
- clone the repository and caffe submodule
- build lib and caffe
- train your model command like:
./experiments/scripts/faster_rcnn_end2end.sh 0 ResNet50 caltech_reasonable
For different datasets, the interval frame rate is different, so it need to be changed.
In ./lib/datasets/caltech.py, change the start_frame and frame_rate of insert_frame function.