Skip to content

Lane Detection in Low-light Conditions Using an Efficient Data Enhancement : Light Conditions Style Transfer (IV 2020)

License

Notifications You must be signed in to change notification settings

Chenzhaowei13/Light-Condition-Style-Transfer

Folders and files

NameName
Last commit message
Last commit date

Latest commit

 

History

49 Commits
 
 
 
 
 
 
 
 
 
 
 
 

Repository files navigation

Light Conditions Style Transfer

Paper

Lane Detection in Low-light Conditions Using an Efficient Data Enhancement : Light Conditions Style Transfer

Accepted by 2020 IEEE Intelligent Vehicles Symposium (IV 2020).

The main framework is as follows: Our framework

Empirically, lane detection model trained using our method demonstrated adaptability in low-light conditions and robustness in complex scenarios. (It can achieve 73.9 F1-measure in CULane testing set)

Datasets

CULane

The whole dataset is available at CULane.

CULane
├── driver_23_30frame       # training&validation
├── driver_161_90frame      # training&validation
├── driver_182_30frame      # training&validation
├── driver_193_90frame      # testing
├── driver_100_30frame      # testing
├── driver_37_30frame       # testing
├── laneseg_label_w16       # labels
└── list                    # list

Generated Images

The images in low-light conditions are generated by the proposed SIM-CycleGAN.

Requirements

  • PyTorch 1.3.0.

  • Matlab (for tools/prob2lines), version R2017a or later.

  • Opencv (for tools/lane_evaluation).

Before start

conda create -n  your_env_name python=3.6
conda activate your_env_name
conda install pytorch==1.3.0 torchvision==0.4.1 cudatoolkit=10.0 -c pytorch
pip install -r requirements.txt 

SIM-CycleGAN

The source code for SIM-CycleGAN has been released. (11/03)

train

Train your own SIM-CycleGAN model as follow.

python train.py  --name repo_name \
                 --dataset_loadtxt_A /path/to/domain_A_txt \
                 --dataset_loadtxt_B /path/to/domain_B_txt \
                 --gpu_ids 6 \

test

Use your trained model to generate images.

python test.py   --name repo_name \
                 --model simcycle_gan \
                 --dataset_loadtxt_A /path/to/domain_A_txt \
                 --dataset_loadtxt_B /path/to/domain_B_txt \
                 --gpu_ids 6 \

Lane Detetcion

The source code used for the lane detction is made publicly available by HOU Yuenan.

Test for Demo

We provide demo for testing a single image or a video.

sh ./demo.sh

You can get the results as follow.

Result for probability map images

Result for points images

If you want to test the model for video, you can set mode=0 in demo.sh.

Evaluate the Model

The trained model used in this paper is available in ./trained.

  1. Run test script
sh ./test_erfnet.sh
  1. Get lines from probability maps
cd tools/prob2lines
matlab -nodisplay -r "main;exit"

Please check the file path in Matlab code before.

  1. Evaluation
cd /tools/lane_evaluation
make
# You may also use cmake instead of make, via:
# mkdir build && cd build && cmake ..
sh eval_all.sh    # evaluate the whole test set
sh eval_split.sh  # evaluate each scenario separately

The evaluation results are saved in /tools/lane_evaluation/output.

Performance

Light Conditions Stlye Transfer

Some examples of real images in normal light conditions and their corresponding translations images in low-light conditions. images

Lane Detetcion

Performance ( (F1-measure) ) of different methods on CULane testing set. For crossroad, only FP is shown.

Category ERFNet CycleGAN+ERFNet SIM-CycleGAN + ERFNet(ours) SCNN ENet-SAD ResNet-101-SAD
Normal 91.5 91.7 91.8 90.6 90.1 90.7
Crowded 71.6 71.5 71.8 69.7 68.8 70.0
Night 67.1 68.9 69.4 66.1 66.0 66.3
No Line 45.1 45.2 46.1 43.4 41.6 43.5
Shadow 71.3 73.1 76.2 66.9 65.9 67.0
Arrow 87.2 87.2 87.8 66.9 65.9 67.0
Dazzle Light 66.0 67.5 66.4 58.5 60.2 59.9
Curve 66.3 69.0 67.1 64.4 65.7 65.7
Crossroad 2199 2402 2346 1990 1998 2052
Total 73.1 73.6 73.9 71.6 70.8 71.8

The probability maps output by the three methods above are shown as following images

To do

  • Add attenction on ERFNet

  • Open the source code for SIM-CycleGAN

  • Upgade pytorch (from 0.3.0 to 1.3.0)

  • Upload demo for test

Citation

Please cite this in your publication if our work helps your research.

@inproceedings{Liu2020Lane,
  title={Lane Detection in Low-light Conditions Using an Efficient Data Enhancement : Light Conditions Style Transfer},
  author={Liu, Tong and Chen, Zhaowei and Yang, Yi and Wu, Zehao and Li, Haowei},
  booktitle={2020 IEEE intelligent vehicles symposium (IV)},
  year={2020},
}

Acknowledgement

This project refers to the following projects: