Details of Grant 

EPSRC Reference:	GR/J10389/01
Title:	A UNIFIED RELAXATION FRAMEWORK FOR INTERPRETING THE DIFFERENTIAL STRUCTURES OF 3D SCENES.
Principal Investigator:	Hancock, Professor E
Other Investigators:
Researcher Co-Investigators:
Project Partners:
Department:	Computer Science
Organisation:	University of York
Scheme:	Standard Research (Pre-FEC)
Starts:	01 October 1993	Ends:	31 March 1997	Value (£):	148,918
EPSRC Research Topic Classifications:	Image & Vision Computing
EPSRC Industrial Sector Classifications:
Related Grants:
Panel History:
Summary on Grant Application Form
This research-programme aims to develop a unified relaxation framework for interpreting the differential structure of 3D scenes represented in terms of density slice data.. Its novelty stems from its adherence to a Bayesian philosophy with different components of the scene model having precise meaning in terms of well defined underlying probability distributions. Conventionally, the interpretation task is approached as a three-stage operation, with different processing steps assigned to the computation of derivatives, the identification of local differential structure and the location of global curves or surfaces. Progress:At the volumetric segmentation level we have developed a novel feature tracking method. It is based on an interframe relaxation technique. This method combines intra-frame and inter-frame constraints on the behaviour of acceptable contour structure. The intra-frame information is represented by both a dictionary of local contour structure and a statistical model of the response of a set of directional feature detection operators. The inter-frame ingredient represents the novel modelling component; it is encapsulated by an implicit model of the underlying surface structure of 3D feature points. The model is represented in terms of a series of unimodal probability densities whose single parameter is the inter-frame distance. The initial probabilities in our relaxation scheme effectively combine distributions describing the statistical uncertainties in the position and feature characteristics of multiframe contours; these probabilities are refined in the light of the dictionary to produce consistent contours. The method has been evaluated experimentally on cranial MRI data. Here the method significantly outperforms its single frame counterpart in term of its ability to extract noise-free and smooth feature contours. Surface fitting and interpretation are tasks of pivotal importance in the analysis of volumetric image data. Although component parts of a well established hierarchy of processing operations, the two processes are inextricably linked to one-another through sharing a common data representation of putative surface points. Conventionally, however, this data is a necessarily condensed and static representation of the available image information. Viewed from this perspective, the flow of information between the two processes has a band limiting on the curvature detail that may be extracted. Moreover, there is little scope for rectifying gross fitting or segmentation errors since there is insufficient evidential information to hand. We have addressed these perceived limitations by casting the processes of surface fitting and differential analysis into a Bayesian framework. We have taken the view that the two processes should be tightly coupled to one-another through a weighted least squares fitting process. The weights are in fact a posteriori probabilities which model both the evidential impact of the associated data and the consistency of the underlying surface points. The coupled fitting process involves iteration between the surface-fiting and surface-analysis modules. Curvature consistency information is used to weight the surface-fit and exclude data non-linearities. Surface fit residuals are used to weight data points in the extraction of consistent curvature information. The net effect is to allow fine curvature detail to be reliably extracted from noisy and discontinuous surfaces. [1] N.G. Sharp and E.R. Hancock, Feature Tracking by Multiframe Relaxation ,to appear in image and Vision Computing, 1995. [2] N.G. Sharp and E.R. Hancock, Feature Tracking by Multiframe Relaxation , Proceedings of the Fifth British Machine Vision Conference: Edited by E.R. Hancock, pp. 407-418, 1994. [3] N.G. Sharp and E.R. Hancock, Multiframe Feature Tracking by Probabilistic Relaxation Shape and Structure in Pattern Recognition: Edited by D. Dori and A. Bruckstein, to appear, 1994. [4] N G Sharp and E R Hanc
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.york.ac.uk