No abstract

Human gesture recognition is defined as a necessity to determine what human body actions occur in videos. Gestures can originate from body motion like walking, bending, jumping, and hand waving. When a video is playing the human action detection is a difficult point of detection. This problem is peculiarly hard due to extensive variations in motion appearance of actions, camera angles with respect to the human body, motion in the background, noise and large amount of video data. Main aim is to detect various gestures in a multimedia clip by preprocessing the video and then applying algorithm for identifying various actions. The important role is to determine behavior of humans based on based on their actions. The moving objects are determined from a video stream. The database used includes CASIA database and WEIZ MANS ACTION database to test the proposed system. Those can be applied to Surveillance Systems.

No abstract