The data used was data provided by Kaggle. The data is largely divided into Violence and NonViolence for 1000 images each.

Real Life Violence Situations Data

I have used InceptionV3 which is a pretrained Imagenet CNN model provided by Keras.

Update - MobileNetV2 is used to with improved accuracy and predictions. This model was created on Kaggle.

Output is in the form of a video, which will tell violence/ non-violence on the top left corner.

https://www.pyimagesearch.com/2019/07/15/video-classification-with-keras-and-deep-learning/ https://www.pyimagesearch.com/2019/06/03/fine-tuning-with-keras-and-deep-learning/