Python script to calculate the rotation speed of the cell from cell videos.

python3
pip install tqdm seaborn scikit-learn scipy Pillow pandas opencv-python opencv-contrib-python numpy matplotlib

0. Get your system ready for the mentioned dependencies.
1. Clone this repo.
2. Inside the script directory, create a new directory named dataset.
3. Copy the cell videos to analyze into the dataset directory.
4. Execute the main script, python main.py.

1. Select ROI for analyzing
   The selection will be cached into python pickle file named roi.pkl, so that you wouldn't need to select the ROI again next time for the same video. Press ENTER to proceed.
2. Adjust the threshold for the binarization
   The optimal threshold gives the most accurate cell area. Press ENTER to proceed.
3. The processing will be started. The results will be outputted into the directory named output.

The output result is a heatmap of pairwise similarity of frame t1 and frame t2. For example,

From this figure, we can observe some peaks in pairwise similarity at delta=30 and delta=60. These peaks indicate the rotation cycles. Given that the rotation cycle is 30 (frames), fps is 30.0, we can calculate that the rotation speed is 360 (deg/s).