An Efficient Activity Detection System based on Skeleton Joints Identification

Abstract

The increasing criminal activities in the current world has drawn lot of interest activity recognition techniques which helps to perform the sophistical analytical operations on human activity and also helps to interface the human and computer interactions. From the existing review analysis it is found that most of the existing systems are not emphasize on computational performance but are more application specific by identifying specific problems. Hence, it is found that all the features are not required for accurate and cost effective human activity detection. Thus, the human skelton action can be considered and presented a simple and accurate process to identify the significant joints only. From the outcomes it is found that the proposed system is cost effective and computational efficient activity recognition technique for human actions.