The Royal Society

Copy of Proceedings in Philosophical Transactions 356(1740) May 1998

Discussion Meeting on
New Geometric Techniques in Computer Vision

Wednesday 23 July and Thursday 24 July 1997
at The Royal Society, 6 Carlton House Terrace, London SW1

Organized by R. Cipolla, J. Lasenby, H.C. Longuet-Higgins and A. Zisserman

Computer vision is the automatic analysis and interpretation of images of real scenes. To date, geometric methods constitute the most useful tools for this analysis. The adoption of projective geometry in the mid-1980's, as a supplement to Euclidean geometry, provided a new approach to many computer vision problems and has led to improved recognition methods and a better understanding of the geometry of multiple views, particularly in the case of uncalibrated cameras. A combination of differential geometry and motion analysis has produced new algorithms for determining surface shape from the temporal evolution of image curves, and a fusion of geometry and statistics has resulted in powerful methods of tracking image features. The meeting will describe recent developments in such techniques, including their implementation in working systems.

Session 1: Geometry of Surfaces
Wednesday 23 July, 09.30-12.30 (Chair: Prof. H.C. Longuet-Higgins)

1. Chairman's introduction
2. Prof. J.J. Koenderink: Pictorial Relief
3. Dr P.J. Giblin: Apparent contours - an outline
4. Dr R. Cipolla - The visual motion of curves and surfaces

Session 2: Structure and Motion
Wednesday 23 July, 14.00-17.40 (Chair: Dr J. Lasenby)

5. Prof. O.D. Faugeras: Grassmann, Cayley and the stratification of 3D vision
6. Prof. T. Kanade: The factorization methods: theory and applications
7. Dr R. Hartley: Projective reconstruction from multiple perspective views
8. Dr A. Zisserman: Metric reconstruction from multiple views

Session 3: Grouping and Matching
Thursday 24 July, 09.30-12.30 (Chair: Dr A. Zisserman)

9. Prof. J.L. Mundy: Object recognition based on geometry: progress over three decades
10. Dr S. Carlsson: Combinatorial geometry for visual recognition
11. Prof. L. VanGool: An invariance based strategy for grouping
12. Dr J. Lasenby: Tensors and 3D-invariants for matching over multiple views

Session 4: Geometry and Statistics
Thursday 24 July, 14.00-17.40 (Chair: Dr R. Cipolla)

13. Prof. A. Blake: Statistical models of visual shape and motion
14. Prof. K. Kanatani: Statistical optimization and geometric inference in Computer Vision
15. Dr P. Torr: Geometric motion segmentation and model selection
16. Prof. J.M. Brady: Summary and closing remarks


For those wishing to attend the meeting there is no cost, but it is required that people register so that numbers are known. Printed programmes and registration forms can be obtained from Clair Emmanuel at the Royal Society (0171-839-5561). Registration must be made by 14th July 1997. If requested, the Royal Society will also send a list of reasonable hotels in the vicinity of the Meeting venue.