PyTorch implementation of MPIIGaze

Requirements

• Python 3.6.3
• PyTorch
• torchvision
• tensorboard-pytorch (optional)
• OpenCV

Download the dataset and preprocess it

$ wget http://datasets.d2.mpi-inf.mpg.de/MPIIGaze/MPIIGaze.tar.gz
$ tar xzvf MPIIGaze.tar.gz

$ python preprocess_data.py --dataset MPIIGaze --outdir data

Usage

$ python -u main.py --arch lenet --dataset data --test_id 0 --outdir results/00

Results

Model	Mean Test Angle Error [degree]	Training Time
LeNet	6.43	3m40s
ResNet-preact-8	5.78	7m27s

$ python -u main.py --arch lenet --dataset data --test_id 0 --outdir results/lenet/00 --batch_size 32 --base_lr 0.01 --momentum 0.9 --nesterov True --weight_decay 1e-4 --epochs 40 --milestones '[30, 35]' --lr_decay 0.1

$ python -u main.py --arch resnet_preact --dataset data --test_id 0 --outdir results/resnet_preact/00 --batch_size 32 --base_lr 0.1 --momentum 0.9 --nesterov True --weight_decay 1e-4 --epochs 40 --milestones '[30, 35]' --lr_decay 0.1

References

• Xucong Zhang and Yusuke Sugano and Mario Fritz and Bulling, Andreas, "Appearance-based Gaze Estimation in the Wild," Proc. of the IEEE Conference on Computer Vision and Pattern Recognition (CVPR), 2015 arXiv:1504.02863, Project Page
• Xucong Zhang and Yusuke Sugano and Mario Fritz and Bulling, Andreas, "MPIIGaze: Real-World Dataset and Deep Appearance-Based Gaze Estimation," arXiv:1711.09017