Deepberry : segmentation and time-lapse tracking of grapevine berries

Installation from source

# for prediction
conda create -n grapevine -c conda-forge python=3.7 tensorflow
conda activate grapevine
conda install -c conda-forge -c numpy pandas scipy tensorflow matplotlib shapely pytest
pip install opencv-contrib-python
pip install pycpd

# (optionnal) for model training (note: for GPU compatibility, see further guidelines at the top of the training scripts): 
pip install -U segmentation-models

# (optionnal) for model validation metrics:
pip install object_detection_metrics

Maintainers

• Benoit Daviet (b.daviet@hotmail.com)
• Fournier Christian (christian.fournier@inrae.fr)