Design And Implementation Of Human Computer Interaction System Based On Features In Face

Human computer interaction (HCI) is an integrated discipline which is merged subjects such as computer application, computer graphics, artificial intelligence and etc. Nowadays, the technology of hardware get a rapid development. And the high demands of interaction from people provide new challenge for HCI technology. Human computer interaction technology, as a burgeoning research in IT, has been widely concerned in many countries. The purpose of HCI is to develop a sort of computer which can see, listen, speak and comprehend the language. Among all kinds of multimedia HCI technology, interaction based on computer vision has a lot of benefits such as high efficiency for interaction, a good experience for users and low cost.The main contents in this paper are as follows:the summary of the background and research actuality of visual interaction technology, detection and location of feature region in face, face to camera distance measurement by monocular vision, precise location of iris and research of interaction algorithm. By fully summarizing the visual interaction systems, a novel interaction method based on feature regions in face is proposed. In order to improve the detection and efficiency rate, the tradition Adaboost algorithm is improved by expanded the Haar features. The face to camera distance is measured by establishing the relationship between distance and pixel in image and constructing the measurement formula. In the iris precise location module, the efficiency and the calculation of Hough transformation is optimized by predicting the radius of iris. By introducing the information of eye’s corners, the point of regard can be judged.In distance measurement module, the distance between face to camera is measured by detecting feature regions in face in monocular vision. Feature regions in face are detected fast by improved AdaBoost algorithm which is ameliorated by sample expanding and feature quantity reducing. By using the system constraints, camera calibration and area mapping, the measurement formula is derived and distance can be measured through that formula. The specific distance experiment verifies the feasible of the system. The complicated background experiment shows that the system performs robust in complicated environment and the accuracy is affected by different illuminations. The system usability experiment verifies the system’s applicability for general users. The accuracy and real-time performance are influenced by glasses which users wear. All the experiment results show that the accuracy is high in the effective range and the real-time requirement can be satisfied.Eyes location is an essential part of human computer interaction. The precise of eyes location decides the feasibility of interaction. In this paper, we expand the features of Haar by introduced a new type of characteristic rectangle which can distinguish the eyes and brow more efficiently. The edge of iris is detected and recovered by coarsely locating eyes, edge extraction and least square fitting. By using the distance information, the radius of iris can be predicted and thus can reduce the calculation of Hough Transformation. By comparing the two improved algorithm, the one which constrained the radius of iris can balance the detection accuracy and calculating efficiency better. This improved algorithm has a good performance on eyes location by increasing the calculation efficiency one order of magnitude.The interaction algorithm based on feature regions in face proposed in this paper has a lot of benefits such as interactive efficiency, usability for different users and adaptability in complicated environment. Besides the whole system need no auxiliary equipment but a camera which means the system is convenient to use and cost less.