This repo contains a project for DSMS using computer vision algorithms.
python3 run_dsms.py
run_dsms.py
takes the following arguments:
-m
: integer value, 1: to use python multiprocessing for make use of all CPU cores; 0: mean do not use multiprocessing.
-np
: integer value equal to the number of parallel processes to use in case of multiprocessing.
-c
: integer value, 1: to run a calibration to find a few parameters for different warnings; 0: do not run calibration.
-sav
: integer value, 1: to save a video clip of the detected driver distraction events; 0: do not save the distraction events.
config.yaml
has many options (see comments in the file) to tune DSMS to optimize accuracy/speed.
You should have OpenCV python installed and working to run this application. Also make sure that you are able to access camera with opencv prior to running this application
opencv-python
imutils
dlib
pyyaml
Documentation for this project is in progress and will be updated here.