Installation

---

After cloning the repo:
cd pheno-wheat
pip install virtualenv (if you don't already have virtualenv installed)
python3 -m virtualenv envwheat to create the virtual environment for the project
source envwheat/bin/activate to activate virtual environment
Use pip install -r requirements.txt to install all the requirements

Additional

1. Follow the CUDA Installation Guide

2. pip install torch==1.13.1+cu{version} torchvision==0.14.1+cu{version} torchaudio==0.13.1 --extra-index-url https://download.pytorch.org/whl/cu{version}
   

3. mim install mmcv==1.6.1
   

Install mmdetection and mmsegmentation by following:

1. mmdetection

git clone https://github.com/open-mmlab/mmdetection.git --branch v2.28.0 --single-branch
cd mmdetection
pip install -v -e .

2. mmsegmentation

git clone https://github.com/open-mmlab/mmsegmentation.git --branch v0.30.0 --single-branch
cd mmsegmentation
pip install -v -e .

Instructions

---

1. Spike Segmentation

Download the Spike segmentation dataset
Extract the zip file in the data directory in the following structure:

pheno-wheat
├── data
│   └── SPIKE_main
│       ├── annotations
│       ├── test
│       ├── train
│       └── val
└── ...

Run bash train_spike.sh to train a model specified in the train_spike.json file under the key config.
(create from train_spike.json.template)

(if permission denied error shows up use the command chmod +x ./train_spike.sh)

Run bash test_spike.sh to test the model specified in the test_spike.json file under the key config loaded from the checkpoint file specified under the key checkpoint.
(create from test_spike.json.template)

To run inference on a single image in demo/ folder, run bash inference.sh.

2. Spikelet Segmentation

Ongoing project