Python scripts for tracking cells in fluorescent microscopy, for Ed Munro's group.
- Python 3.6
- Python scientific stack (
numpy,scipy,matplotlib) scikit-imagepykalmanffmpegoravconvcommand line tools for creating movies
Everything is pure python except for the function write_movie in BKlib.py, which uses ffmpeg/avconv. One of these will probably be installed on Linux machines; it can be installed on Mac using the binaries here. The scripts should work on Windows, except for the function write_movie, which depends on avconv.
All of the scripts take a tiff movie as input. You can see help instructions by typing the script name without any arguments. I've included a sample movie, cell_membranes.tiff.
track_cell.py must be run first. It takes a tiff file as input, does a simple watershed-based segmentation, and presents a GUI to the user to select a region of interest (i.e., a single segmented cell). This region is then tracked across subsequent frames using an active contour/snakes model. Finally, the control points from the contours are saved to an npy file for futher analysis. This output npy file is the main output; take note of its name, since it is a required argument for the other scripts.
visualize_tracking.py shows the movie with the tracking results overlaid. Use the -o flag to specify an output avi filename in order to save the visualization as a movie.
create_cell_movie.py extracts a small window around the cell of interest and saves it to a movie, so that the motion of the cell is "frozen". Note that the tiff can be a different channel from that used for cell tracking, i.e., you can use a channel that displays cell boundaries in track_cell.py, but another channel with this script.
plot_cell_trajectory.py computes the centroids of the cell boundaries, smooths the trajectory with a Kalman filter, and creates a few plots to quantify motion.