Bilkent University
Department of Computer Engineering
MS Thesis Presentation
Model-Based Camera Tracking for Augmented Reality
Aytek Aman
MS Student
Computer Engineering Department
Bilkent University
Augmented reality (AR) is the enhancement of real scenes with virtual entities. It is used to enhance user experience and interaction in various ways. Educational applications, architectural visualizations, military training scenarios and pure entertainment-based applications are often enhanced by augmented reality to provide more immersive experience. With hand-held devices getting more powerful and cheap, such applications are becoming very popular.
To provide natural AR experiences, camera parameters must be calculated in an accurate and robust way so that virtual entities can be overlaid onto the real environments correctly. Estimating camera parameters in real-time is a challenging topic. In most systems, visual tracking serve as the main method for estimating the camera pose. For rich-textured environments, keypoint-based methods work quite well. Edge-based tracking, on the other hand, is more preferable when the environment is rich in geometry but has little or no visible texture.
Edge-based trackers fail in particular scenarios, especially in the presence of edge mismatches between the projected model and the video feed. We propose a number of methods to improve the quality and performance of edge-based tracking. We preprocess view-dependent edge reliability and visibility to improve the edge-based tracking process. Additionally, we use more accurate adaptive projection algorithm to provide more uniform control point distribution in the screen space.
We test our camera tracker in different environments to show the effectiveness and performance of the proposed algorithms. The preprocessed visibility enables constant time calculations for edge visibility while preserving the accuracy of the tracker. Additionally, the reliability scores calculated per control point make tracking more accurate under camera configurations where the edge mismatch rate is potentially high.
Finally, we demonstrate a sample AR application with user interaction to present our AR framework, which is developed for a commercially available and widely used game engine.
DATE: 28 August, 2014, Thursday @ 14:00
PLACE: EA-409