The purpose of this project is to classify the gender of a Three Spined Stickleback based on automatic landmarks detection and maching learning classification
10 Specific landmarks are computed through Machine Learning (regression trees) according to Kazemi method adapted for biological research [1,2,3,4] :
[1] Kazemi,Sullivan, "One millisecond face alignment with an ensemble of regression trees," 2014
[2] Perrot,Bourdon,Helbert "Implementing cascaded regression tree-based face landmarking" 2020
[3] Porto, Voje "ML-morph: A fast, accurate and general approach for automated detection and landmarking of biological structures in images" 2020
[4] Irani, Allada.. "Highly versatile facial landmarks detection models using ensemble of regression trees with application" 2019
Once points are detected, collection of 46 distances are calculated : body length, head length, jaw length...
Finally, a machine learning model predicts the gender of the Three Spined Stickleback :
No dependencies required, everything is inside the exe file. Just launch the GUI
- python==3.8.6
- sklearn==0.23.2
- numpy==1.18.5
- scipy==1.5.2
- dlib==19.22.0
- xgboost==1.4.2