This code is associated with the paper from Hol et al., "BiteOscope, an open platform to study mosquito biting behavior". eLife, 2020. http://doi.org/10.7554/eLife.56829
Cages can be made out of acrylic using a laser cutter, design files are provided in the cageDesigns directory. Temperature control is provided by tempControl.py and requires a raspberry pi, a waterproof DS18B20 temperature probe, and a 5V relay.
Image analysis code is provided as .py files and notebooks (notebook filenames are appended with 'NB') and can be tested using the demo data available as a zip file.
- trackMosq.py finds centroids of all mosquitoes in images and tracks them over time.
- trackingResultsMovie.ipynb creates a video in which tracking results are marked to verify output.
- cropTracks_features.py stores cropped images centered on the focal mosquito of all frames belonging to a single track and calculates various features (e.g. movement and feeding stats).
- inferenceAlbo_test.py does DeepLabCut based body part tracking
The playground folder contains various notebooks for downstream analysis and is under continuous development.
The biteOscope code uses the following modules:
- numpy
- matplotlib
- pandas
- scipy
- scikit-image
- scikit-learn
- trackpy
- opencv
- joblib
- deeplabcut
- tqdm