Anchor-frame-based Efficient Authoring for Mobile Augmented Reality

Abstract

Augmented reality has been used increasingly due to the development of mobile devices. One of the important features in augmented reality is the authoring that makes virtual contents in the real world. Thus, many companies have released their own software development kits for the authoring. In this thesis, the approach that make the authoring efficiently using anchor-frame for mobile augmented reality is discussed.

Simultaneous localization and mapping (SLAM) is adopted as a tracking method by analyzing other tools. Among the existing SLAM methods, a hybrid method that uses both sensors and vision information is selected for stability. However, depending on the type of SLAM, the location of the object on the map is changed. Therefore, anchor-frame is defined for fixing an object in the correct position. The experiment shows that the anchor-frame makes the position of the virtual object stable.

In addition, we analyzed the interaction method for convenient authoring tool through user study. Experiments were conducted in three stages: Generation, Initial Pose, and Manipulation. Generation is to decide what and where to create. In this experiment, hand and voice were compared and there was no difference in usability. Initial Pose is to change the position on the screen to a three-dimensional position. Different methods were good depending on the situation. Manipulation is to change the state of the created object. It was better to manipulate the values on three axes through addition interface in the same position. An efficient authoring tool can be created using these experimental results.