BeAwareOfYourAct: A Framework for Behavioural Action Detection in Workplace through Deep Learning Analysis and Augmented Action Pattern Recognition
Edward Humphrey C. Isles, Francis F. Balahadia · 2022
The purpose of this study is to propose a system framework that addresses the problem in monitoring malicious behavior of an individual towards someone in the same vicinity through multi-person detection video processing. The real-time video rendering analysis and behavioral action detection is an efficient method in line with pervasive or any event related atrocious behavior towards an individual. In the development of the system several algorithms will be utilized such as object detection, action recognition algorithms and techniques (e.g. Human Activity Recognition, Supervised Learning Data Representation, etc.) which will heavily rely on Regions with Convolutional Neural Network (RCNN). The capability of Artificial Intelligence (A.I.) to analyze real-time video rendering through the training data to be supplied as a point of reference in deep learning analysis, any act that might cause discomfort to the victim can be detected to be reported immediately to the Officer-in-charge preventing the maltreatment to further escalate. In order to accomplish the said study that will serve as a pillar for the system development. It is concluded that this system framework of real-time monitoring is a great approach in preventing certain abuse from starting and will make the conduct of investigation more precise and efficient should a harassment come into place. Through this framework it can make the vicinity more conducive to healthy communication and prevent intimidating words to be casually mentioned.