Artem Evdokimov Supervisor: Prof. Gudrun Klinker Advisor: Rudolph, Linda (@ge29tuw) Submission Date: [created]
Today, augmented reality is gradually gaining popularity in various applications. Modern approaches based on the usage of a 2D color camera make it possible to visualise virtual objects in the real-world 3D space with a sufficiently high accuracy. Therefore, virtual objects can appear to stay in the same place in relation to the real world, even if a camera is tilted to look at the object from different perspectives. However, some imperfections caused by different types of scenes captured by the camera can still occur, which makes tracking quality to be dependent on lighting conditions, types of surfaces and other images quality degradation factors. This thesis proposes an approach to notify a user regarding the areas of the surrounding real-world 3D space that are difficult to track, so that even users without prior knowledge in the field of augmented reality can adjust the conditions of a tracked space to stabilize natural feature tracking and improve the overall quality of AR based applications.
[ PDF (optional) ]
[ Slides Kickoff/Final (optional)]