yangwonkyu / kitti

Kitti on Faster R-CNN (Python implementation)

Contents

1. Requirements
2. Basic installation
3. Usage
4. Results
5. Evaluation
6. GUI

If you have any questions with this repository, please feel free to contact me at [e-mail]

Requirements

1. Requirements for Caffe and pycaffe (see: Caffe installation instructions)

Note: Caffe must be built with support for Python layers!

# In your Makefile.config, make sure to have this line uncommented
WITH_PYTHON_LAYER := 1
# Unrelatedly, it's also recommended that you use CUDNN
USE_CUDNN := 1

2. Python packages you might not have: cython, python-opencv, easydict
3. [Optional] MATLAB is required for official PASCAL VOC evaluation only. The code now includes unofficial Python evaluation code.

Installation

1. Clone the Kitti repository

# Make sure to clone with --recursive
git clone --recursive https://github.com/czhu95/kitti.git
cd kitti
git checkout master

2. We'll call the directory that you cloned KITTI_ROOT

   Ignore notes 1 and 2 if you followed step 1 above.

   Note 1: If you didn't clone KITTI with the --recursive flag, then you'll need to manually clone the caffe-fast-rcnn submodule:

   git submodule update --init --recursive

   Note 2: The caffe-fast-rcnn submodule needs to be on the faster-rcnn branch (or equivalent detached state). This will happen automatically if you followed step 1 instructions.

3. Build the Cython modules

   cd $KITTI_ROOT/lib
   make

4. Build Caffe and pycaffe

   cd $KITTI_ROOT/caffe-fast-rcnn
   # Now follow the Caffe installation instructions here:
   #   http://caffe.berkeleyvision.org/installation.html
   
   # If you're experienced with Caffe and have all of the requirements installed
   # and your Makefile.config in place, then simply do:
   make -j8 && make pycaffe

5. Download pre-computed Faster R-CNN detectors

   cd $KITTI_ROOT
   ./data/scripts/fetch_faster_rcnn_models.sh

   Please download the pre-computed kitti model here:
   http://pan.baidu.com/s/1dEZOXOl password: w8n4
   And place the .caffenet model file under KITTI_ROOT/data/kitti_models/

6. Create symlinks for the kitti dataset (Not necessary if you wish only to run test)

     cd $KITTI_ROOT/data
     ln -s $kitti kitti

   Please make sure the kitti dataset has this basic structure

     $kitti/                         # kitti dataset
     $kitti/image                    # holds all the images
     $kitti/image/train              # training images
     #kitti/label                    # text annotations

Usage

To train a kitti vehicle detector.

cd $KITTI_ROOT
mkdir output
./experiments/scripts/train_kitti.sh [GPU_ID] 
# GPU_ID is the GPU you want to train on

Output will be in KITTI_ROOT/output/

To test a kitti vehicle detector.

cd $KITTI_ROOT
mkdir results
./experiments/scripts/test_kitti.sh [GPU_ID] [TEST_DIR]
# GPU_ID is the GPU you want to train on
# TEST_DIR is the directory containing test images (default to data/kitti/image/test)

Output will be in TEST_DIR/label. If TEST_DIR is not specified, output will be stored at KITTI_ROOT/results/

Results

We ran our detector on a test image set given by TA. If you wish to evaluate the results, please download here:
http://pan.baidu.com/s/1hsc0Fzu password: 4pny
Note: These results do not include class specification. If you want the alternative, please re-run the test procedure as described above.

Evaluation

We implemented matlab code to draw PR curve given test text files. Please refer to ./evaluation for codes and furthur instructions.
Note: This implementation is for results with classifications.

GUI

We wrote a matlab gui program to display detection results. Please refer to ./gui for codes and furthur instructions.
Note: This implementation is for results with classifications.