Implementation of Faster RCNN by Tensorflow (In development)
・Complete
Load Images of KiTTI Object Detection Datasets
Preprocessing for Network Input
RPN(Region Proposal Network)
Proposal Layer(Convert rpn to rois)
・ToDO
Trainer for RCNN
# Prepare KiTTI Datasets
http://www.cvlibs.net/datasets/kitti/eval_object.php
# Compile Cython File
cd cython_util
./setup.sh
# Training RPN
cd rpn
python rpn.py
ROI Pooling layer was implemented by this repository
https://github.com/deepsense-io/roi-pooling