Skip to content

Landmark detection approach for the Pepeketua ID project.

License

Notifications You must be signed in to change notification settings

wildlifeai/pepeketua_landmarks

Folders and files

NameName
Last commit message
Last commit date

Latest commit

 

History

28 Commits
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 

Repository files navigation

Landmark detection - Pepeketua ID

Pepeketua ID is an open-source pattern recognition software that facilitates the individual identification of New Zealand endemic frogs (Pepeketua).

"(Overview of the three main modules and the components of the Koster Seafloor Observatory.")

You can find out more about the project at https://www.wildlife.ai/projects/pepeketua-id/

Overview

This repository contains scripts related to the Landmark Detection component of the Pepeketua ID.

The Landmark Detection model predicts six frog landmarks in an image. These landmarks will be used to morph and crop the image to facilitate the individual identification of each frog.

"(Example of the frog landmarks predicted by the model.")

Example of the landmarks (red points) produced by the model.

Quickstart

Requirements

  • Python 3.7+
  • Python dependencies listed in requirements.txt

Instructions

Download the repo, and after building the docker with build_docker.sh, run run_docker.sh <dir_path> where dir path is a path to your dir with all the images. It will automaticly find all images in sub dirs as well.

Citation

If you use this code or its models in your research, please cite:

Hay G, Carmon E, Vinograd B, Anton V (2021). An open-source landmark detection model approach to facilitate the individual identification of New Zealand endemic frogs. https://github.com/wildlifeai/pepeketua_landmarks

Collaborations/questions

We are working to make our work available to other herpetologits. Please feel free to contact us with your questions.

About

Landmark detection approach for the Pepeketua ID project.

Resources

License

Stars

Watchers

Forks

Contributors 3

  •  
  •  
  •