Towards ROXAS AI: Deep learning for faster and more accurate conifer cell analysis

by Marc Katzenmaier, Vivien Sainte Fare Garnot, Jesper Björklund, Loïc Schneider, Jan Dirk Wegner, Georg von Arx

[Paper] [Dataset] [BibTex]

Exemplary impovement over ROXAS

Install

create a python environment and run the following command to install all requirements

pip install torch torchvision opencv-python numpy pytorch-lightning segmentation-models-pytorch scikit-image glob2 

Usage

Download the model from here and run the model with

python run_model.py --input=./input/001.jpg --output=./output/001.png

This code will save the output in a .png file. The png files are for visual inspection. If the model is not in the TowardsRoxasAI folder specify the path with --model=path_to_model

Note: This script will run the model only on the CPU which is significantly slower than on the GPU. This is done to make it as accessible as possible for everyone to try out if the model will work on his data.

Dataset

The dataset can be downloaded form Zenodo

Contributors

We thank A. Ivanova, A. Arzac, A. Piermattei, A.L. Prendin, A. Kaiser, A. Louy, D. Castagneri, D. Benito, D. Noordermeer, G. Petit, G. Giberti, H. Song, I. Serra Olabuenaga, K. Janecka, K. Seftigen, L. Petrucco, L. Mateju, L. Haberbauer, M. Ferriz Nunez, M. Guerin, M. Klisz, M. Carrer, M. Tabakova, M. Fonti, M. Vos, M. Dell'Oro, M. Rydval, N. Maredova, P. Fonti, R. aus der Au, R. Peters, V. Shishov, T. Pampuch and V. Simanko for providing training images for the deep learning models.

Cite

If you use Towards Roxas AI in your research, please use the following BibTeX entry.

@article{KATZENMAIER2023126126,
    title = {Towards ROXAS AI: Deep learning for faster and more accurate conifer cell analysis},
    author = {Marc Katzenmaier and Vivien Sainte Fare Garnot and Jesper Björklund and Loïc Schneider and Jan Dirk Wegner and Georg {von Arx}}
    journal = {Dendrochronologia},
    year = {2023},
    doi = {https://doi.org/10.1016/j.dendro.2023.126126},
}