Sponsored by
YouArt
issai / fes-face
Model's Last Updated: July 28 2026
object-detection
Introduction of fes-face
Model Details of fes-face
FES Face Bounding Box Detection (Event Cameras)
Environment Installation & Set up
To launch training or inference, first of all it is required to install Metavision SDK environment.
-
If you use Linux OS, install Metavision SDK environment from this link .
-
If you use Windows OS, install Metavision SDK environment from this link .
Git repo:
https://github.com/IS2AI/faces-in-event-streams
The following scripts can be found withhin our github repo:
Face detection model
To run inference for face bounding box model, use the following command:
python3 detection_and_tracking_pipeline.py --object_detector_dir /path/to/model --record_file <RAW file to process> --display
alternatively you can proceed with instructions from this link.
Before launching training, please place label_map_dictionary.json, which comes as part of this repo, in the same folder where your train/val/test folders are located.
- To train the model to detect face bounding box in event streams, run:
cd <path to train_detection.py>
python3 train_detection.py <path to output directory> <path to dataset>
alternatively, you can follow instructions from here .
- To select the feature extractor, you need to define --feature_extractor option using:
python3 train_detection.py <path to output directory> <path to dataset> --feature_extractor <select feature extractor: Vanilla, ResNet_18, ResNet_34, ResNet_50>
Facial landmark detection model
To train or detect five point facial landmark detection, the change to the following files should be done:
- feature_extractors.py
- box_processing.py
- display_frame.py
- modules.py
Changes which need to be done can be found in our repo in files:
- box_processing_difference.py
- display_frame_difference.py
- modules_difference.py
- feature_extractors_difference.py
Further, train and inference instructions are the same as for face detection part, declared above.
Since
- feature_extractors.py
- box_processing.py
- display_frame.py
- modules.py
are files which originally comes as part of Metavision SDK which comes under Prophesee Metavision Licensing, we will share only changes that has to be done on the installed Metavision SDK package (once you install Metavision SDK, you will have direct access to the original files).
If you use the dataset/source code/pre-trained models in your research, please cite our work:
@Article{s24051409,
AUTHOR = {Bissarinova, Ulzhan and Rakhimzhanova, Tomiris and Kenzhebalin, Daulet and Varol, Huseyin Atakan},
TITLE = {Faces in Event Streams (FES): An Annotated Face Dataset for Event Cameras},
JOURNAL = {Sensors},
VOLUME = {24},
YEAR = {2024},
NUMBER = {5},
ARTICLE-NUMBER = {1409},
URL = {https://www.mdpi.com/1424-8220/24/5/1409},
ISSN = {1424-8220},
DOI = {10.3390/s24051409}
}
Runs of issai fes-face on huggingface.co
0
Total runs
0
24-hour runs
0
3-day runs
0
7-day runs
0
30-day runs
More Information About fes-face huggingface.co Model
Url of fes-face
fes-face huggingface.co Url
Provider of fes-face huggingface.co
Other API from issai
Total runs:
319
Run Growth:
49
Growth Rate:
15.36%
Total runs:
298
Run Growth:
51
Growth Rate:
17.11%
Total runs:
146
Run Growth:
-15
Growth Rate:
-10.27%
Total runs:
46
Run Growth:
1
Growth Rate:
2.17%
Total runs:
43
Run Growth:
43
Growth Rate:
100.00%
Total runs:
14
Run Growth:
14
Growth Rate:
100.00%
Total runs:
11
Run Growth:
-3
Growth Rate:
-27.27%
Total runs:
8
Run Growth:
0
Growth Rate:
0.00%
Total runs:
7
Run Growth:
7
Growth Rate:
100.00%
Total runs:
6
Run Growth:
6
Growth Rate:
100.00%
Total runs:
6
Run Growth:
0
Growth Rate:
0.00%
Total runs:
0
Run Growth:
-4
Growth Rate:
0.00%
Total runs:
0
Run Growth:
-11
Growth Rate:
0.00%
Total runs:
0
Run Growth:
0
Growth Rate:
0.00%