A Support System for Collaborative Creative Activities based on Augmented Reality Technology

1.多用户协同交互技术的发展视频透视增强现实系统存在虚拟对象有时会在真实的图像中隐藏重要用户活动的问题。我们开发了一种方法来绘制虚拟对象与正确的遮挡关系，通过测量从相机到真实的对象的实时深度。此外，我们还开发了一种交互技术，通过检测用户在图像中通过手势或指向工具指向的位置来操作虚拟对象或改变系统模式。2.开发了一种技术来管理/表示具有物理属性的多个想法我们开发了软件来表示桌子上方、桌子上或手持显示设备上的各种信息或虚拟对象。此外，我们还开发了一个手势接口来控制分布式增强现实系统共享的虚拟对象。此外，我们还研究了阴影表示和透明表示对虚拟对象的影响。3.基于特征学习的对象跟踪技术的发展我们提出了一种方法，可以根据从一组点特征和线特征中自动提取的特征的跟踪结果，稳定地计算出真实的对象的姿态和位置。我们还对该方法进行了改进，同时使用了灰度图像模板和边缘图像模板，使其能够在特征较少的情况下很好地工作。此外，我们提高了跟踪的鲁棒性，通过管理多个跟踪假设。

1.Development of collaborative interaction techniques by multiple usersVideo see-through augmented reality system has a problem that virtual objects sometimes conceal an important user's activity in the real image. We developed a method for drawing virtual objects with the correct occlusion relationship by measuring the depth from a camera to the real objects in real-time. Also we developed an interaction technique to operate virtual objects or to change system modes by detecting the place where the user is pointing by gesture or pointing tools in the image.2.Development of a technology to manage/represent multiple ideas that are given physical propertiesWe developed software to represent various information or virtual objects above a table, on the table or a handheld display device. Also we developed a gesture interface to control virtual objects that are shared by distributed augmented reality systems. In addition, we examined the effect of shadow representation and transparent representation for virtual objects.3.Development of object tracking technology based on feature learningWe developed a method to calculate the pose and position of a real object stably from the tracking results of features that are automatically extracted from a set of point features and line features. We also improved the method by using both intensity image templates and edge image templates so that it can work well in the environment that there are few features. Also we improved the robustness of the tracking by managing multiple tracking hypotheses.