Awesome
š Retina : Low-Power Eye Tracking with Event Camera and Spiking Hardware š
š» Blog |š Paper | šļø Data
Retina : Low-Power Eye Tracking with Event Camera and Spiking Hardware
š§š»āš Pietro Bonazzi <sup>1</sup>,
Sizhen Bian <sup>1</sup>,
Giovanni Lippolis <sup>2</sup>,
Yawei Li<sup>1</sup>,
Sadique Sheik <sup>2</sup>,
Michele Magno<sup>1</sup> <br>
<sup>1</sup> ETH Zurich, Switzerland <br> <sup>2</sup> SynSense AG, Switzerland
Quick Comparaison: Ground Truth (GT) vs. Prediction
In the following GIFs, Yellow represents the Ground Truth (GT), and Green represents the Prediction. These images are taken from the validation set.
<table> <tr> <td><img src="docs/0_val_3_0.gif" width="250" height="250" /></td> <td><img src="docs/0_val_12_0.gif" width="250" height="250" /></td> <td><img src="docs/0_val_16_0.gif" width="250" height="250" /></td> <td><img src="docs/0_val_19_0.gif" width="250" height="250" /></td> </tr> </table>āļø Citation ā¤ļø
Leave a star to support our open source initiative!āļø
@InProceedings{Bonazzi_2024_CVPR,
author = {Bonazzi, Pietro and Bian, Sizhen and Lippolis, Giovanni and Li, Yawei and Sheik, Sadique and Magno, Michele},
title = {Retina : Low-Power Eye Tracking with Event Camera and Spiking Hardware},
booktitle = {Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR) Workshops},
month = {June},
year = {2024},
pages = {5684-5692}
}
š TL;DR quickstart š
Clone the repo
git clone https://gitlab.ethz.ch/pbonazzi/retina.git
cd retina
Create the environment
conda create -n retina python=3.10 numpy=1.8.1
conda activate retina
pip install -r requirements.txt
pip install git+https://gitlab.com/inivation/dv/dv-processing.git
Downloads
Click here to download the dataset.
Verify the structure:
.
āāā name
ā āāā annotations.csv
ā āāā events.aedat4
āāā ...
āāā silver.csv
Click here to download a pretrained model.
Training
See the list of arguments in the launch_fire function. The run-name has the format version-name.
python train.py --run-name="1-train"