Repository files navigation

RON is a state-of-the-art visual object detection system for efficient object detection framework.The code is modified frompy-faster-rcnn.You can use the code to train/evaluate a network for object detection task.For more details, please refer to ourCVPR paper.

If you find RON useful in your research, please consider citing:

@inproceedings{KongtCVPR2017,    Author = {Tao Kong, Fuchun Sun, Anbang Yao, Huaping Liu, Ming Lu, Yurong Chen},    Title = {RON: Reverse Connection with Objectness Prior Networks for Object Detection},    Booktitle = {IEEE Conference on Computer Vision and Pattern Recognition},    Year = {2017}}

Note: SSD300 and SSD500 are the original SSD model fromSSD.

0. Clone the RON repository

   git clone https://github.com/taokong/RON.git

1. Build Caffe and pycaffe

   cd $RON_ROOT/git clone https://github.com/taokong/caffe-ron.gitcd caffe-ronmake -j8 && make pycaffe*this version use CUDNN for efficiency, so make sure that "USE_CUDNN := 1" in the Makefile.config file.

2. Build the Cython modules

   cd $RON_ROOT/libmake

3. installation for training and testing models on PASCAL VOC dataset

   3.0 The PASCAL VOC dataset has the basic structure:

    $VOCdevkit/                           # development kit $VOCdevkit/VOCcode/                   # VOC utility code $VOCdevkit/VOC2007                    # image sets, annotations, etc.

   3.1 Create symlinks for the PASCAL VOC dataset

    cd $RON_ROOT/data ln -s $VOCdevkit VOCdevkit2007 ln -s $VOCdevkit VOCdevkit2012

4. Test with PASCAL VOC dataset

   Now we provide two models for testing the pascal voc 2007 test dataset. To use demo you need to download the pretrained RON model, please download the model manually fromBaiduYun(Google Drive), and put it under$data/RON_models.

   4.0 The original model as introduced in the RON paper:

    ./test_voc07.sh # The final result of the model should be 74.2% mAP.

   4.1 A lite model we make some optimization after the original one:

    ./test_voc07_reduced.sh # The final result of the model should be 74.1% mAP.

5. Train with PASCAL VOC dataset

   Please download ImageNet-pre-trained VGG models manually fromBaiduYun(Google Drive), and put them into$data/ImageNet_models. Then everything is done, you could train your own model.

5.0 The original model as introduced in the RON paper:     ./train_voc.sh    5.1 A lite model we make some optimization after the original one:    ./train_voc_reduced.sh