Main Article Content
Activity recognition is one of the significant task and most difficult tasks in computer science. Several actions are performed by the various people and the activities are collected based on the actions performed by the people. Many existing approaches are developed to recognize the several individual actions, pair-wise interactions, and pose which is not an easy task. Deep Learning (DL) is most widely used to solve the various issues in human activity recognition. In this paper, the Enhanced algorithm for human activity recognition is developed with the integration of Generative Adversarial Network (GAN) with HaaR features with bounding box. This is a very efficient method that will provide the individual activities and also the group activities, actions that are recognized by the proposed approach. The performance is calculated by using parameters such as accuracy. The comparative results are CERN, CAR, and Enhanced GAN.