3D model-based tracking combining edges, keypoints and fiducial markers

Abstract

Model-based tracking is an essential task in fields such as Augmented Reality. State-of-the-art approaches rely on the model’s edges, sometimes combined with image keypoints and color. Nevertheless, these image features are not considered part of the model but as temporary information discarded every time the tracking process is restarted. This paper proposes a novel approach that employs an enhanced model that combines edges, keypoints, and fiducial markers for robust and real-time tracking. The experiments conducted show that our method outperforms state-of-the-art modelbased approaches and suggest that fiducial markers are a good choice for texturing models.