mathiramilo / pupil-detection

Python pupil detection in human eye images project for the Computer Image Processing course of the university of the republic UDELAR.

Computer Image Processing Final Project - Pupil Detection

Pupil detection in human eye images using image processing techniques with python.

Website

Learn more about this project by visiting it's website in the following URL:

https://mathiasramilo.pages.fing.edu.uy/proyecto-final-timag/

Usage

On Linux:

• Install the required packages:

pip3 install -r requirements.txt

• Segment pupil on a single image:

python cli_runner.py --image_path <path_to_image>

Note: omitting the --image_path argument will run the program on a default image.

• Evaluate the model over a folder of images:

python cli_runner.py --folder_path <path_to_folder>

Note: omitting the --folder_path argument will run the program on a default folder.

• Evaluate the model over a dataset:

python cli_runner.py --dataset_path <path_to_dataset>

Note: omitting the --dataset_path argument will run the program on a default dataset. Moreover it will run over the LPW dataset.

Dataset

• Dataset Website
• Paper Link

Review of State of the Art Eye Gaze Estimation Methods

D. W. Hansen and Q. Ji, "In the Eye of the Beholder: A Survey of Models for Eyes and Gaze," in IEEE Transactions on Pattern Analysis and Machine Intelligence, vol. 32, no. 3, pp. 478-500, March 2010, doi: 10.1109/TPAMI.2009.30.

Primary Source

This paper served as our primary source, aiming not only to reference its findings but also to faithfully reproduce the results it achieved.

• Citation