Title: STag: A Stable Fiducial Marker System

Abstract: Fiducial markers provide better-defined features than the ones naturally available in the scene. For this reason, they are widely utilized in computer vision applications where reliable pose estimation is required. Factors such as imaging noise and subtle changes in illumination induce jitter on the estimated pose. Jitter impairs robustness in vision and robotics applications, and deteriorates the sense of presence and immersion in AR/VR applications. In this paper, we propose STag, a fiducial marker system that provides stable pose estimation. STag is designed to be robust against jitter factors, thus sustains pose stability better than the existing solutions. This is achieved by utilizing geometric features that can be localized more repeatably. The outer square border of the marker is used for detection and homography estimation. This is followed by a novel homography refinement step using the inner circular border. After refinement, the pose can be estimated stably and robustly across viewing conditions. These features are demonstrated with a comprehensive set of experiments, including comparisons with the state of the art fiducial marker systems.