Detecting Harmful Activity in Pilgrimage Using Deep Learning

Abstract

CCTV surveillance is the most extensively used intelligent latest innovation. The use of surveillance cameras has risen dramatically because of the convenience of monitoring from anywhere and the reduction of crime rates in public areas.  In this paper, we introduce the idea of bad vibe activity detection from live videos to enhance the security and safety of pilgrims.  The proposed bad vibes activity recognition model is intended to be addressed in the most efficient manner possible using cutting-edge technologies such as TensorFlow and Keras.   TensorFlow was chosen because the project could be deployed to a mobile environment in the future with the possibility of extension of other areas such as airport security, bus stain, and public areas that may deserve special attention for security checks. We choose MediaPipe Holistic for employee bad vibe recognition in the model.