Deep learning-based methods for individual recognition in small birds

This repository provides the necessary guidance for the implementation of our paper ”Deep learning-based methods for individual recognition in small birds”, including a step-by-step on how to automatize the collection of training pictures using RFID technology, pre-process the obtained images and train convolutional neural networks for individual identification. The tutorials presented in this repository are meant for researchers with very basic programing skills (preferentially python language).

Citation and Contact

@article{deeplearningbirds,
  title={Deep learning-based methods for individual recognition in small birds},
  author={André C. Ferreira, Liliana R. Silva, Francesco Renna, Hanja B. Brandl, Julien P. Renoult, Damien R. Farine, Rita Covas and Claire Doutrelant},
  journal={Methods in Ecology and Evolution},
  year={2020}
}

For more information please contact andremcferreira@cibio.up.pt or andrem343@hotmail.com

Table of Contents

1. Requirements

2. Automated_pictures_collection: raspberry pi communication with RFID data logger to build a camera trap with automatic labelling.

3. Data_pre-processing: localize and extract birds from pictures, pictures transformation and similarity index.

4. Train_CNN: training the convolutional neural network for individual identification

5. Dataset