mmbl-track

Results are stored by name of input folder of the png stack. Compact results has the particle id, total displacement, total number of frames it was tracked across, direction list, frame list of where it was tracked and simple displacement. Total displacement is displacement along the specific path that the particle took (inaccurate sometimes due to oscillation detected back and forth). Simple displacement is the distance between the final and intial point for the swimmer. For each tracked particle, detailed results includes a frame by frame snapshot of the dx, dy, total displacement and change in direction.

Setup

This repo is written in Python 2. You may need to install pip2.7 to install packages in Python 2 on your machine:

$ wget https://bootstrap.pypa.io/pip/2.7/get-pip.py
$ sudo python2.7 get-pip.py
$ which pip2.7

If $ which pip2.7 returns a directory, then everything has installed correctly. Now you can use the following steps to set up a virtual environment with the correct packages for this repo.

If you are in an OS environment you may need to install wget using % brew install wget.

1. Create a virtual environment using:

$ python -m virtualenv ./env
$ source env/bin/activate

2. Install the required packages at the correct versions using:

$ pip2.7 install -r requirements.txt

To get a list of the required packages, run: $ pip2.7 freeze

And copy the results into the requirements.txt file. Make sure to do this inside of an active virtualenv in Python2!

Walkthrough

Sample pipeline for analysis. Assume video files are grouped by amplitude for each sample in a test protocol. We will be analyzing test protocol a for this walkthrough. For example, media/tpa/s1/a100 has multiple mp4 files inside it. All commands need to be run in the root directory where the scripts are located.

1. python createFolders.py --input media/tpa

This puts all of the videos found recursively in the media/tpa folder into their own folders and renames them for later scripts. NOTE: The video files in this directory must be .mp4 format.

2. python createPNGStack.py --input media/tpa/a100

This converts all of the videos located in media/tpa/s1/a100 to a png stack. This command needs to be run individuall on all of the amplitude subfolders (ex. media/tpa/s1/a75)

3. python blurAll.py --input media/tpa/

For blurring the videos and removing the excess markers. Creates a cloned folder with a png stack in same root directory of original folder containing png stack. For example media/tpa/s1/a100/s1_a100_f1_v1/ will now have a blurred clone at media/tpa/s1/a100/s1_a100_f1_v1_blurred.

4. mkdir results/tpa

Make the results directory

5. python fullAnalysis.py --input media/tpa/ --output results/tpa

Run the analysis on the entire test protocol. Note, will put all results folder in results/tpa so might be helpful to run fullAnalysis on individual samples ie. python fullAnalysis.py --input media/tpa/s1/ --output results/tpa/s1 for more organized files

6. python graphAll.py --input results/tpa

Create plots from results and store them in the results directory

User-Defined Values

createPNGStack.py

1. Line 35, modify fps value (do not make higher than sampling fps)

fullAnalysis.py

1. Line 46, startFrame for video analysis. Defaulted to 180 allowing time for setup

main.py

1. Line 53, two key values are diameter and minmass, use getParameters jupyter notebook to calibrate for sample.
2. Line 56, can modify maxChange and memory to filter velocities
3. Line 170, can turn labels on or off on the trajectory plot

graph.py

1. Line 21-23 minDist, fps, scalar. (What is scalar?)