Research On User Visual Attention Distribution Characteristics Oriented To The 3D Gesture Interface

People tend to express their ideas through different gestures in human’s daily communication, which makes gesture play a very important role in life. Therefore, Gesture interaction has been a hot issue in the field of human-computer interaction and researchers have been working on more natural, harmonious interactive algorithm with cognitive behavior. But the current algorithms are all tracking information at each frame and the amount of calculation is very big, therefore, these algorithms lead to inefficient interaction. As a result, we have to reflect on the question that based on the gesture not accurately under the condition of direct interaction, whether it is necessary for every frame gesture tracking and obtain the mathematical model for 3 D gesture.This thesis is supported by the National Natural Science Foundation of China(Grant Nos. 61173079 and 61472163) and Key Project of Natural Science Foundation of Shandong Province(Grant No. ZR2011FZ003). The study of gesture interaction algorithm have already been a long time and made many achievements, but the bottlenecks of tracking accuracy and low efficiency of interaction is still difficult to break through. By studying the distribution of visual attention in the user interaction process, this thesis could try to break through the bottlenecks of tracking accuracy in existing methods. The visual attention mechanism is a very complex process because of its unusual complexity and uncertainty, which covers many disciplines of cognitive science, neurobiology, psychology etc. The human visual system always screen out interested user objects quickly, ignore or abandon the other object which is not interested.This suggests that we can make the allocation of resources quickly and reasonable, thus improves the work efficiency of the visual processing system in the process of handling information. In this thesis, main tasks include the following two aspects. On the one hand, the main task is to study the distribution of human visual attention in this thesis and visual attention mechanism has the ability to filter data information which provides the basis for improving the efficiency of gesture interaction, so it has important theoretical significance. On the other hand, at present, because the 3D real-time tracking of gestures is very difficult in general PC conditions, so this thesis research from a new perspective to explore the real-time hand gesture tracking, which has important application significance in the design of new 3D in human-computer interaction interface.Main research work and innovations of this article are as follows:(1) Tobii eye tracker was used to realize real-time tracking of the human eye sightIn order to study the distribution law of visual attention, people’s line of sight data recorded by using Tobii eye tracker. The Tobii eye tracker can record the most accurate data in the most natural user state, which is very important to study the distribution of visual attention. For gesture interaction experiment by using the Tobii eye tracker, we can see the user visual attention time in each region and thus can learn the user’s area of interest and not the interested region.(2) The establishment of visual attention modelDigging into the eye movement data, visual attention model are established for translation, grab, release stage in gesture interaction process. These models are studied in-depth from the perspective of visual attention strength and the distance between gesture model and target location.Proved by experiments, visual attention distribution completely accords with human cognitive behavior in gesture interaction.(3) 3D gesture tracking algorithm based on the visual attention distribution is presentedThe algorithm innovation is combined with particle filter and visual attention mechanism and the particle is adaptive in the improved algorithm. This algorithm not only solves the bottleneck problem of gesture interaction algorithm, also accords with the natural habit of people. In simple terms, with the change of visual attention intensity, the particles are also different and visual attention intensity is lower, then we use animation to achieve. By experimental verification, the algorithm can improve the gesture tracking accuracy and interaction efficiency and the interaction accords with people’s natural interaction behavior, which also reflects new human computer interaction with "people" as the center.The human visual attention mechanism is able to screening of data information rapidly. This thesis improves the efficiency of gesture interaction through the applications of visual attention mechanism to gesture interaction which break a new path for the study of the human-computer interaction. Explore the real-time gesture tracking from the perspective of visual attention mechanism also provides a new method for the application of gesture tracking which has a certain popularization value.