Details of Grant 

EPSRC Reference:	GR/J15681/01
Title:	REAL TIME TRACKING OF SURFACES IN MOTION
Principal Investigator:	Blake, Professor A
Other Investigators:
Researcher Co-Investigators:
Project Partners:
Department:	Engineering Science
Organisation:	University of Oxford
Scheme:	Standard Research (Pre-FEC)
Starts:	01 June 1993	Ends:	30 November 1995	Value (£):	76,223
EPSRC Research Topic Classifications:	Image & Vision Computing
EPSRC Industrial Sector Classifications:
Related Grants:
Panel History:
Summary on Grant Application Form
1. To develop flexible parametric surface models for tracking with automatic control of spatio temporal scale. 2. To develop automatic aquisition of models, learning variability both of shape and of non-rigid motion.3. To demonstrate tracking of biological forms with the Oxford Range Sensor, in particular to track a flexing hand in real time.Progress:This research has been based around the Oxford range sensor. This projects a set of light stripes simultaneously onto the scene, so that range measurements can be computed from just a single image frame. This allows moving objects to be scanned by effectively freezing their motion. A complete range image is computed in a few seconds in software.Research into tracking deofrmable objects using the sensor has fallen into the following key areas:1. The work has centred around using deformable surface models to track the position and shape of a target as it moves under the sensor. The initial models have consisted of parametric surface patches, augmented with uncertainty models describing the position and velocity of the surface. Using predictive filtering, objects can be tracked at frame rate as they move under the sensor at speeds of up to 10cm/s. 2. In order to achieve real time performance, it is necessary to select just a few range measurements to make at each each step. Within the predictive filter framework, the theory of optimal measurements has been developed. Based on the current uncertainty of the model, the best set of measurements to make can be chosen (in the sense that these will maximally increase the information of the model) [1]. 3. The surface has been initially modeled as a tensor product B-spline. This class of surface is sufficient to track continuous regions, such as the back of a hand, but to include fingers in the model it would he desirable to introduce tears. To do this, the standard tensor product formulation has been extended, allowing the introduction of extra degrees of freedom and different topologys into the surface description.The final stages of the research are being concentrated on the efficient initialisation of the object models onto their targets, and the comparison of linear versus non-linear models used for tracking.[1] P.J. Lindsey and A. Blake. Real-time tracking of surfaces with structured light. In BMVC 94, pages 619-28, 1994. Also in Image and Vision Computing, in press, 1994.
Key Findings
This information can now be found on Gateway to Research (GtR) http://gtr.rcuk.ac.uk
Potential use in non-academic contexts
This information can now be found on Gateway to Research (GtR) http://gtr.rcuk.ac.uk
Impacts
Description This information can now be found on Gateway to Research (GtR) http://gtr.rcuk.ac.uk Summary Date Materialised
Sectors submitted by the Researcher
This information can now be found on Gateway to Research (GtR) http://gtr.rcuk.ac.uk
Project URL:
Further Information:
Organisation Website:	http://www.ox.ac.uk