Hand Gesture Recognition

With the development of science and technology, people are constantly proposing new demands and challenges on the quality and level of interactivity. In view of intuitive and the natural characteristics of hand gesture, it has become an important means of information exchange between the people and machines. Vision-based dynamic hand gesture recognition system is currently one of the hot field of scientific research, many scholars at home and abroad have invested a great deal of enthusiasm for research, while vision-based gesture recognition system is the main development trends in the current and future periods.This study is of non-contact hand gesture recognition system based on the single-camera gesture, including five modules research which are image preprocessing, hand gestures segmentation and positioning, fingertip detection, trajectory feature extraction and gesture recognition, and gives my own design.Main results are as follows:Firstly, the thesis studies the background and significance of gesture recognition, gesture recognition and the development was reviewed at home and abroad.Secondly, the paper provides a brief overview on the basic knowledge about image preprocessing and hand gesture segmentation and positioning technology, and realize a better gesture segmentation making use of H, S components of skin color detection and location based on HSV space combining the dynamic nature.Thirdly, the fingertip in the writing process in the air has always been the highest point, this paper use this feature to detect the fingertip's rough location, and then realize the accurate positioning of fingertip with the ring characteristics of the fingertip. After it,we can record its location and extract the characteristics of the fingertip trajectory.Lastly, Because Hidden Markov Model has a good ability of modeling the timing signal, this thesis selects the Hidden Markov Model (HMM) to train the trajectory characteristics and to achieve trajectory recognition. Finally, our results demonstrate that making use of HMM to realize the hand gesture trajectory recognition can get better recognition results.