Abstract
In this paper we present an algorithm for real-time object classification and human activity recognition which can help to made intelligent video surveillance systems for human behavior analysis. The proposed method makes use of object silhouettes to classify objects and activity of humans present in a scene monitored by a dynamic camera. An statical background subtraction method is used for object segmentation. The matching templates are constructed using the motion history images for classify objects into classes like human, human group and vehicle; and object shape information for different human activities in a video. Experimental results demonstrate that the proposed method can recognize these activities accurately.
Published Version
Talk to us
Join us for a 30 min session where you can share your feedback and ask us any queries you have
Schedule a call