Harris corner detector and blob analysis featuers in human activty recognetion

The automated detection and monitoring of human activities have gained increased attention in the last decade due to many video applications. They are playing a central role of behavior analysis of human being, where adequate monitoring can minimize the risk of harm to our society. Although, the activities recognition has been studied by many researchers but it still inaccurate. This because of high similarity between human joints when its move to perform some activities such as walking, running and jogging. In this paper, a human activity recognition system was designed based on features extraction analysis. Two types of features extractions techniques were used, which are the basic blob analysis features and Harris corner detector. By comparing the accuracy of the recognition rate in each technique through the two scenarios we found that Harris corner detector is more powerful than the basic blob analysis features because of it is capable to distinguish between the similar activities in an accurate manner.