| With rapid development of computer technology, new software and hardware applications are constantly emerging, in contrast to which the "one mouse-a computer" model of human-computer interaction (HCI) has been maintained for many years without changes. The classic interaction can't take up more complex human-machine interaction. MultiTouch, a new technology concept, is a new multi-point input and multi-point response technology, which has greatly improved the environment for HCI. With this technology, the key point of HCI has been transferred from computer to human, which directs the future trend of HCI.In this paper, a realization of MultiTouch technology programs was put forward according to the systematic function requirement. In testing of multi-point input, single-purpose computer vision technology was applied in this paper by using two detection terminals to detect two gesture input. In visual detection and identification of gestures, the Harr-like feature of the gestures which could enhance the speed of detection was collected at first. Then the Harr-like features of the input gestures was processed by Cascade classifier and its algorithm, with which the classifier had auto attention function and could achieve high detection rate and low rate of false identification. At last, based on the requirement of identifying two gestures with high speed, a specific gesture detection and identification method was discussed in the paper. In processing the multi-input information, two functions were used to simulate multithreading, based on which the processing method was designed. This program also has a good expansibility and can easily be extended to more than two input application.A practical system has been developed successfully based on the technology put forward in this paper. It has been proved that the system has reached the requirement and could run steadily. |
