Object Pose, Shape and State Estimation for Augmented Reality

Object-centered Augmented Reality is of importance for a wide area of applications such as industrial maintenance and quality assurance, education, entertainment and many others. Early challenges for this area included the detection, 6DoF pose estimation and tracking of objects using geometric computer vision or machine learning methods. Presently, new problems are in focus such as dealing with particularly challenging objects, modelling hand-object interactions, including objects in global mapping, or performing category-level object pose estimation.

The main goal of this Research Topic is to showcase new technologies in object pose, shape and state estimation that will shape Augmented Reality in the following years. There are particular elements that create the promise of even more
compelling AR experiences with increased practical value and applicability. These technologies include precise localization of objects under all conditions, improved interaction capabilities leading
to a tight coupling between the virtual and real world as well as full semantic understanding of environments at a level of object relations, dependencies and functionalities.

For this Research Topic, authors are invited to submit original work on novel approaches, state-of-the-art review surveys or innovative applications with relation to the topic of object-based AR in the
following areas:

• Object pose estimation and tracking
• Category-level object shape
• Hand and object pose and shape estimation
• Object-level SLAM systems
• Object illumination for AR
• Machine Learning/Deep Learning object-focused tasks
• 3D Model to real object domain adaptation
• Depth sensing and sensor fusion
• Reflective, feature-less and transparent object understanding
• Object-oriented AR applications