| Summary: | 碩士 === 中華大學 === 電機工程學系碩士班 === 87 === With the development of multimedia system and the environment
of virtual reality, the communication between human and PC
becomes a highlighting issue. Therefore, the development of
HCI(Human-Computer-Interface) becomes more and more important.
Gesture is one of the best natural ways to communicate. In
this thesis, we proposed real time and efficient methods,
dynamic time warping and radial basis function neural network,
to recognize the hand gestures. Traditionally, the technology
of gesture recognition was divided into two categories:
vision-based and glove-based methods. Vision-based methods has
been popularly used in some researches and applications.
Generally, computer camera is the input device for observing
the information of hands for fingers. However, the computation
complexity in tracking of hands has several bottlenecks, such
as feature extraction, objects need separated from background,
fingers motion tracking, etc. Thus, it is difficult to achieve
real time operation. For these reasons, we have turned to
glove-based technique which is more feasible and more practical
in gesture recognition.
In this thesis, we have developed an intelligent and user-
friendly recognition system based on low-cost personal computer
, on-line learning, high recognition rate and real time
operation. The resulting system is designed to be more robust,
users may employ previous gestures or defining gestures by
themselves in order to satisfy different requirements.
|
|---|
