pupildetector
Science Score: 26.0%
This score indicates how likely this project is to be science-related based on various indicators:
-
○CITATION.cff file
-
✓codemeta.json file
Found codemeta.json file -
✓.zenodo.json file
Found .zenodo.json file -
○DOI references
-
○Academic publication links
-
○Academic email domains
-
○Institutional organization owner
-
○JOSS paper metadata
-
○Scientific vocabulary similarity
Low similarity (10.1%) to scientific vocabulary
Repository
Basic Info
- Host: GitHub
- Owner: Guyazulay2
- License: other
- Default Branch: master
- Size: 189 MB
Statistics
- Stars: 0
- Watchers: 0
- Forks: 0
- Open Issues: 0
- Releases: 0
Metadata Files
README.md
PupilDetector
This project uses the DeepVOG model, presented in the credits section below. The original DeepVOG is a framework for pupil segmentation and gaze estimation based on a fully convolutional neural network. Currently, it is available for offline gaze estimation of eye-tracking video clips.
In this work, we used the DeepVOG model to predict the pupil center coordinates of a given eye image and NOT a video as in the original DeepVOG project. Also, we allow this prediction by handling HTTP requests that contain input eye images.
Installation by a docker image
Pull the docker image to run the Flask server:
ruby
$ docker pull 247555/pupildetector:latest
Usage (Command-line interface)
Note: The following 'fit' and the 'infer' commands are only available through the DeepVOG repository, mentioned in the credits section. Also, these functionalities don't depend on running the Flask server as described in 2. below.
In order to use the DeepVOG model, it is necessary to first fit an eyeball model on a given input video. There is an example of that as a JSON file available in the repository. The deserved command is:
ruby
$ python -m deepvog --fit [VIDEO_FIT_PATH] [EYEBALL_MODEL_PATH]
If there is no need to fit an eyeball model as described, there are two ways to predict the pupil center:
1. By the "infer" mode:
ruby
$ python -m deepvog --infer [IMAGE_PATH] [EXISTING_EYEBALL_MODEL_PATH] [CSV_RESULTS_FILE_PATH] -v [IMAGE_INFERENCE_PATH]
- By an HTTP request
2.1 Run the Flask server through the container:
ruby $ docker run -d --runtime=nvidia -p 5000:5000 247555/pupildetector:latest python /notebooks/DeepVOG_dspip/deepvog/_main_.py
2.2 Send the HTTP request:
ruby
$ curl -X POST -F "image_file=@exmp.jpg" http://$PUBLIC_IP:5000/
Credits
For more information, visit the original DeepVOG repository in the following link: https://github.com/pydsgz/DeepVOG
Owner
- Name: GuyAzulay
- Login: Guyazulay2
- Kind: user
- Repositories: 1
- Profile: https://github.com/Guyazulay2
GitHub Events
Total
Last Year
Dependencies
- keras >=2.2.4
- numpy >=1.12
- scikit-image >=0.14.0
- scikit-video >=1.1.0
- tensorflow-gpu >=1.12.0
- keras >=2.2.4
- numpy >=1.12
- scikit-image >=0.14.0
- scikit-video >=1.1.0
- tensorflow-gpu >=1.12.0