Abstract: This paper presents a technique for human object detection and recognition that uses basic model for back ground subtraction. Background subtraction approach is used to detect the moving object from background. Our proposed technique takes the first frame from a live video (from web cam/CCTV) when it is switched on and compares it with the next frames of the video sequence in an interval. Thus when it detects an object system analyzes it. If it is a human being alarm rings. The system used in the study ranges from varying levels of accuracy and computational complexity. In future enhancement of this work is remotely we are checking and providing security to our system.

Keywords: Background subtraction, Real-time systems, Video surveillance, Segmentation, Human recognition, Alarm system.