Gesture Recognition Using Web Camera

Gesture recognition represents an important ability by which a computer is able to directly accept human gesture as input to trigger different actions just like conventional input devices such as keyboard. mouse, joystick and etc. As the Human Computer Interaction (HCI) progresses over the years, emphasis is placed more on developing input devices which are most convenient and easy to use. Human gesture is not only natural and intuitive to a user, but can also represent motions of high degree of freedom which is of utmost importance in many applications especially in virtual reality. This thesis presents the design of an offline system which is capable of recognizing hand postures from the visual input of a web camera. The hand segmentation is based on image subtraction technique and a skin color modeling process. Fourier descriptors are used as the features to describe the geometry of different hand postures while the recognition process is based on minimum distance classifier. The results obtained indicate that the system is able to recognize hand postures with reasonable accuracy

Text FK_2004_25 IR.pdf Download (2MB)

Actions (login required)

View Item