position_tracking

An assortment of jupyter notebooks to track rodent position using deeplabcut.

Created to use deeplabcut to track animal position.

Pipeline:

Run on local machine within DLC environment. You can find detailed instructions for setting up DLC here.
After running the notebook, you must save a copy of the project folder and video folder (videos to analyze) to your google drive and update paths in the config.yaml file.

Open Colab_TrainNetwork_VideoAnalysis.ipynb in colab to make use of their GPUs.

Training network and analyzing videos may take some time. Colab GPUs disconnect after extended use. Current version saves every 500 iterations. Training can resume at last saved iteration if GPU disconnects.

Use Check_DLC.ipynb to verify tracking accuracy for each video. If tracking accuracy is poor, additional images can be labeled and network can be retrained to improve accuracy.

Notebooks are not general. Paths or DLC parameters should be adjusted as needed.

Name		Name	Last commit message	Last commit date
.gitignore		.gitignore
Check_DLC.ipynb		Check_DLC.ipynb
Colab_TrainNetwork_VideoAnalysis.ipynb		Colab_TrainNetwork_VideoAnalysis.ipynb
README.md		README.md
create_avi_from_bitfields_testing.ipynb		create_avi_from_bitfields_testing.ipynb
create_avi_from_neuralynx_bitfields.ipynb		create_avi_from_neuralynx_bitfields.ipynb
initialize_project_app_ps1.ipynb		initialize_project_app_ps1.ipynb
initialize_project_bitfields.ipynb		initialize_project_bitfields.ipynb