This module prerform violence detection in video using 3-D CNN(Convolution Neural Netowork) and LSTM(Long Short Term Memory)
This combination of 3D CNN and LSTM used as 3D CNN extract features form the frames of the video with spatial details and LSTM layer stores important feature/data over the sequence of frame over the span of vide
This model also can be used to perform violence in live video using CV2 which was included in final part of the python note book above
Flask backend was created to perform the violence detection by uploading the video html css and js used for front end
Refer the video link for the working demo [https://drive.google.com/file/d/1xbUukItsyBgU_WLvxHIYvvhUI_N13ySS/view?usp=sharing]