Itracker

The Pytorch Implementation of "Eye Tracking for Everyone". (updated in 2021/04/28)

We build benchmarks for gaze estimation in our survey "Appearance-based Gaze Estimation With Deep Learning: A Review and Benchmark". This is the implemented code of the "Itrack" method in our benchmark. Please refer our survey for more details.

We recommend you to use data processing codes provided in GazeHub. You can direct run the method' code using the processed dataset.

Links to gaze estimation codes.

• A Coarse-to-fine Adaptive Network for Appearance-based Gaze Estimation, AAAI 2020 (Coming soon)
• Gaze360: Physically Unconstrained Gaze Estimation in the Wild, ICCV 2019
• Appearance-Based Gaze Estimation Using Dilated-Convolutions, ACCV 2019
• Appearance-Based Gaze Estimation via Evaluation-Guided Asymmetric Regression, ECCV 2018
• RT-GENE: Real-Time Eye Gaze Estimation in Natural Environments, ECCV 2018
• MPIIGaze: Real-World Dataset and Deep Appearance-Based Gaze Estimation, TPAMI 2017
• It’s written all over your face: Full-face appearance-based gaze estimation, CVPRW 2017
• Eye Tracking for Everyone, CVPR 2016
• Appearance-Based Gaze Estimation in the Wild, CVPR 2015

Performance

The method is evaluated in three tasks. Please refer our survey for more details.

License

The code is under the license of CC BY-NC-SA 4.0 license.

Introduction

Gazecapture dataset provides training, test and validation set. We provide code for the evaluation with training-test mode. In addition, cross validation is also a common evaluation strategy in gaze estimation. We also provide code for the evaluation with cross validation mode.

The project contains following files/folders.

• model.py, the model code.
• train.py, the entry for training with training-test mode.
• test.py, the entry for testing with training-test mode.
• leave_train.py, the entry for training with cross-validation mode.
• leave_test.py, the entry for training with cross-validation mode.
• config/, this folder contains the config of the experiment in each dataset. To run our code, you should write your own config.yaml.
• reader/, the code for reading data. You can use the provided reader or write your own reader.
• run.sh, a shell file for cross validation.

Getting Started

Writing your own config.yaml

Normally, for training, you should change

1. train.save.save_path, The model is saved in the $save_path$/checkpoint/.
2. train.data.image, This is the path of image, please use the provided data processing code in GazeHub
3. train.data.label, This is the path of label.
4. reader, This indicates the used reader. It is the filename in reader folder, e.g., reader/reader_mpii.py ==> reader: reader_mpii.

For test, you should change

1. test.load.load_path, it is usually the same as train.save.save_path. The test result is saved in $load_path$/evaluation/.
2. test.data.image, it is usually the same as train.data.image.
3. test.data.label, it is usually the same as train.data.label.

Training

In the cross-validation mode, you can run

python leave_train.py config/config_mpii.yaml 0

This means the code will run with config_mpii.yaml and use the 0th person as the test set.

You also can run

bash run.sh leave_train.py config/config_mpii.yaml

This means the code will perform cross validation automatically.
run.sh performs iteration, you can change the iteration times in run.sh for different datasets, e.g., set the iteration times as 4 for four-fold validation.

In the training-test mode, you can run

python train.py config/config_gazecapture.yaml

Test

In the cross-validation mode, you can run

python leave_test.py config/config_mpii.yaml 0

or

bash run.sh leave_test.py config/config_mpii.yaml

In the training-test mode, you can run

python test.py config/config_gazecapture.yaml

Result

After training or test, you can find the result from the $save_path$ in the config file.

Citation

If you use our code, please cite:

@InProceedings{Krafka_2016_CVPR,
	author = {Krafka, Kyle and Khosla, Aditya and Kellnhofer, Petr and Kannan, Harini and Bhandarkar, Suchendra and Matusik, Wojciech and Torralba, Antonio},
	title = {Eye Tracking for Everyone},
	booktitle = {The IEEE Conference on Computer Vision and Pattern Recognition (CVPR)},
	month = {June},
	year = {2016}
}


@article{Cheng2021Survey,
        title={Appearance-based Gaze Estimation With Deep Learning: A Review and Benchmark},
        author={Yihua Cheng and Haofei Wang and Yiwei Bao and Feng Lu},
        journal={arXiv preprint arXiv:2104.12668},
        year={2021}
}

Contact

Please email any questions or comments to yihua_c@buaa.edu.cn.

Reference

1. MPIIGaze: Real-World Dataset and Deep Appearance-Based Gaze Estimation
2. EYEDIAP Database: Data Description and Gaze Tracking Evaluation Benchmarks
3. Learning-by-Synthesis for Appearance-based 3D Gaze Estimation
4. Gaze360: Physically Unconstrained Gaze Estimation in the Wild
5. ETH-XGaze: A Large Scale Dataset for Gaze Estimation under Extreme Head Pose and Gaze Variation
6. Appearance-Based Gaze Estimation in the Wild
7. Appearance-Based Gaze Estimation Using Dilated-Convolutions
8. RT-GENE: Real-Time Eye Gaze Estimation in Natural Environments
9. It’s written all over your face: Full-face appearance-based gaze estimation
10. A Coarse-to-fine Adaptive Network for Appearance-based Gaze Estimation
11. Eye Tracking for Everyone
12. Adaptive Feature Fusion Network for Gaze Tracking in Mobile Tablets
13. On-Device Few-Shot Personalization for Real-Time Gaze Estimation
14. A Generalized and Robust Method Towards Practical Gaze Estimation on Smart Phone