Static Hand Gesture Recognition And Its Application Based On Hu Moments & Support Vector Machine

Recently, more and more Human-Computer Interaction systems are making use of the interface based on Hand Gesture Recognition. There are two methods on hand gesture recognition, recognition based on data glove and recognition based on vision. Taking hand gesture as the input equipment directly, the gesture recognition based on vision is more natural, and it can embody the human vision simulation of the computer, so it has attracted increasing attention of the researchers.This paper deals with static hand gesture recognition based on vision and proposes a new approach based on Hu Moments and Support Vector Machine. Then the approach is applied in two situations, one is the simulation of the hand gesture recognition and those hand gestures stand for zero to nine. In addition, the other applies it to the hand completeness checking in the Driver Physical Examination System.First, foundational theories about the common approaches of image pretreatment, geometric moment and Support Vector Machine are introduced. The uncertainty of rotation, scale of hand gesture brings many difficulties to the extraction of feature. Geometric Moment is an arithmetic based on statistics. This article applies the arithmetic in the extraction of gesture feature since the feature can remain the same when the image is rotated and scaled. SVM is a novel learning method based on statistical learning theory, possessing academic foundation and excellent learning ability, and having a lot of issue in machine learning area. It is based on VC dimension theory, adopting the SRM principle and special advantages in dealing with small samples, non-linear pattern recognition in high dimension.The second half part of this paper focused on the application of the Static Gesture Recognition based on Hu Moments and SVM. Test shows that the combination of the two theories gains high recognition rate that is proved 98.7% in number gesture recognition simulation. Moreover, the application of software for the hand completeness in Driver Physical Examination System is satisfactory.This paper puts forward a new algorithm for static hand gesture recognition based on Hu moments and SVM according to their specific characteristics. This is the first innovation of the article, and the second is that the proposed algorithm is successfully introduced to the Driver Physical Examination System. The results of the simulation of number gesture recognition and its application in Driver Physical Examination System show that the approach based on Hu Moments and SVM deserves more attention and more research in static hand gesture recognition.