vFI-features

Code for running features for the visual FI model There are 5 python scripts to run the different vFI features

Requirements

Developed in Python 3.6 and Python 3.7 Many of these scripts require pose files generated from the Kumar Lab's mouse pose estimation neural network. Contact us for more information.

Generating Pose, Gait, Ellipse fit

*Pose

*Gait

*Ellipse fit

Rear Paw Widths

rearpawwidths.py will produce the median and mean rear paw width features. It requires the following packages:

h5py
os
statistics
pandas
argparse
csv
re

Running it requires 3 arguments:

--input-root-dir: the root directory for input files
--output-root-dir: the root directory for output files (defaults to same as input dir)
--video-file-list: the csv list of videos to process (default is all AVI files in the input dir). Pose files of the videos should be in the same directory as the videos

Example code for command line:

python /home/user/code/rearpawwidths.py --input-root-dir /home/user/data --output-root-dir /home/user/results --video-file-list /home/user/mousefiles.csv

Flexibility Index

flexindex.py will produce the spinal mobility measures (dAC, dB, aABC).

It requires the following packages:

h5py
os
statistics
pandas
argparse
csv
re

Running it requires 3 arguments:

--input-root-dir: the root directory for input files
--output-root-dir: the root directory for output files (defaults to same as input dir)
--video-file-list: the csv list of videos to process (default is all AVI files in the input dir). Pose files of the videos should be in the same directory as the videos

Example code for command line:

python /home/user/code/flexindex.py --input-root-dir /home/user/data --output-root-dir /home/user/results --video-file-list /home/user/mousefiles.csv

Rearing

rearingmetrics.py will produce the rearing features.

It requires the following packages:

h5py
os
statistics
pandas
argparse
csv
re
imageio
numpy
cv2

Running it requires 3 arguments:

--input-root-dir: the root directory for input files
--output-root-dir: the root directory for output files (defaults to same as input dir)
--video-file-list: the csv list of videos to process (default is all AVI files in the input dir). Pose files of the videos should be in the same directory as the videos

Example code for command line:

python /home/user/rearingmetrics.py --input-root-dir /home/user/data --output-root-dir /home/user/results --video-file-list /home/user/mousefiles.csv

Elipse Fit

ellipsefit.py will produce the medain/mean widths and lengths. It does not require pose files but it requires npy files with the width and length of the ellipse fit at each frame. Contact us for more information. It requires the following packages:

os
pandas
argparse
csv
re
nupmy
statistics

Running it requires 3 arguments:

--input-root-dir: the root directory for input files
--output-root-dir: the root directory for output files (defaults to same as input dir)
--video-file-list: the csv list of videos to process (default is all AVI files in the input dir). Pose files of the videos should be in the same directory as the videos

Example code for command line:

python /home/user/ellispefit.py --input-root-dir /home/user/data --output-root-dir /home/user/results --video-file-list /home/user/mousefiles.csv