Corner-based Region Proposal Network

CRPN is a two-stage detection framework for multi-oriented scene text. The code is modified from py-faster-rcnn.

Requirements

Clone the CRPN repository

git clone https://github.com/xhzdeng/crpn.git

Build Caffe and pycaffe (see: Caffe installation instructions)
```
cd caffe-fast-rcnn
make -j8 && make pycaffe
```
Build the Cython modules
```
cd $CRPN_ROOT/lib
make
```
Prepare your own training data. It should have the basic structure followed PASCAL VOC dataset
```
Create symlinks for YOUR dataset
cd $CRPN_ROOT/data
ln -s [dataset_path] VOCdevkit
```

Download pre-trained ImageNet VGG-16 models

You can find it at [Caffe Model Zoo](https://github.com/BVLC/caffe/wiki/Model-Zoo)

Train with YOUR dataset

cd $CRPN_ROOT
./experiments/scripts/train.sh [NET] [MODEL] [DATASET] [ITER_NUM]
# NET is the network arch to use, only {vgg16} in this implemention
# MODEL is the pre-trained model you want to use to initial your weights
# DATASET points to your dataset
# IETR_NUM

Test with YOUR dataset

cd $CRPN_ROOT
./experiments/scripts/test.sh [NET] [MODEL] [DATASET]
# NET is the network arch to use, only {vgg16} in this implemention
# MODEL is the resulting model you trained before
# DATASET points to your dataset

Test outputs are saved under:

output/<experiment directory>/<dataset name>/<network snapshot name>/

xijunjun/crpn

Corner-based Region Proposal Network

Requirements