Abnormal event detection is a growing demand to process a plethora of surveillance videos. Our project helps detect the abnormality in videos with high accuracy, thus saving time for organizations and individuals who would have to go through the entire footage instead.
Python 38.59%OpenEdge ABL 61.41%
abnormal-event-detection's Introduction
Pre-Required software
1. Python 2.7x
2. Modules
- opencv
- numpy
Running steps:
:::::::::Training:::::::::
1 -> Copy the folder containing the frames of the video to be tested to location "/frames/Train/Train1" or "/frames/Train/Train2"
2 -> Go to the folder which has "project.py"
3 -> Run it using the command "python project.py"
4 -> In training mode option type "reload" to train using the frames provided in "/frames/Train"
:::::::::Testing:::::::::
1-> Copy the folder containing the frames of the test video to be tested to location "/frames/Test/Test1"
2 -> Go to the folder which has "project.py"
3 -> Run it using the command "python project.py"
4 -> In training mode option type anything other than reload
5 -> It takes the chached copy of learned feature to test the new frames provided in "/frames/Test/Test1"
:::::::::Ouput:::::::::
1 -> After testing it'll prompt the user to press enter to watch the frames which are abnormal
2 -> After watching 1 can be pressed to reload the frames and watch again any number of times or any other key to test the next video
:::::::::::::::::::::::::::