This is the source code for the machine learning based version of the drosophila Appendometer. It performs landmark prediction in drosophila images using the ml-morph pipeline.
- Clone the repo:
git clone https://github.com/agporto/Appendometer && cd Appendometer/
- Create a clean virtual environment
conda create -n append python=3.7
conda activate append
- Install dependencies
python -m pip install --upgrade pip
pip install -r requirements.txt
Once the packages are installed, you can use the codebase by running the appendML.py command line interface. This file takes in several optional arguments, which are described below:
-ior--input-dir: (optional) The input directory (default = images)-por--predictor: (optional) The trained shape prediction model (default = resources/predictor.dat)-oor--out-file: (optional) The output filename suffix (default = output.xml)-lor--ignore-list: (optional) prevents landmarks of choice from being output
Example prompt:
python appendML.py -i images/ -p resources/predictor.dat
The resulting output.xml file can be used to visualize the predicted landmarks in imglab.
Note that appendML.py will automatically sort images files according to their inferred prediction error. The images with the highest error will be shown first.
Coming soon!