Category: Augmented Reality
The realistic interaction of real and synthetic content is an important part of creating a believable augmented reality. This augmentation may be as simple as rendering markers as if they are actually attached to physical objects, or as complex as adding 3D special effects to live video, but in either case the synthetic geometry must appear fixed to the same frame as the objects around it in order to be believable. The key to rendering geometry in a fixed frame is to maintain an accurate estimate of the camera position relative to the environment. Purely IMU-based solutions suffer from inaccuracy, and particularly lag, in the estimated position which results in synthetic geometry which appears to float over the real geometry. More fundamentally, however, the fact that no analysis is carried out on the captured image stream means that the system cannot know what is actually within the field of view of the camera. Thus, the system cannot take into account line of sight and must render synthetic content based only on viewpoint direction, rather than on the basis of the immediately visible environment. ACVT has undertaken a large body of fundamental research in this field and has several active projects looking at many different applications of the technology.
ACVT has been awarded a 3 year ARC Linkage Grant valued at $282,217. The industry partner on the grant is Canon Information Systems Research Australia. CiSRA is the Australian R&D Centre for Canon Inc., a world leader in professional business and consumer imaging equipment and information systems. The CIs on the project are Prof. Anton van […]
Anton van den Hengel recently attended ISMAR 2009 in Orlando, Florida, presenting; In Situ Image-based Modeling by Anton van den Hengel, Rhys Hill, Ben Ward, and Anthony Dick.
ACVT has been awarded a 3 year ARC Discovery Grant valued at $225,000. The CIs are Prof. Anton van den Hengel in collaboration with Dr Philip Torr of Oxford-Brookes University in the UK. Many industry sectors are becoming increasingly reliant on the production of 3D models from imagery. These models are used for purposes such […]
VideoTrace is a system for interactively generating realistic 3D models of objects from video—models that might be inserted into a video game, a simulation environment, or another video sequence. The user interacts with VideoTrace by tracing the shape of the object to be modelled over one or more frames of the video. By interpreting the […]