Output details
12 - Aeronautical, Mechanical, Chemical and Manufacturing Engineering
University of Leeds
Shape detection with vision: implementing shape grammars in conceptual design
The problem of sub-shape detection prevents application of shape grammars to design synthesis problems involving arbitrary shapes in visual objects. The international shape grammar community has spent over 30 years searching for analytical solutions without success. This paper reports a novel application of object recognition from computer vision to shape grammar implementation. Hausdorff distance is used as a similarity measure of arbitrary shapes, resulting in the world's first computational design synthesis system that works with arbitrary emergent shapes in designers’ 2D sketches. The research led to a further project (Leverhulme Ref: F/00269/Z) and an EU proposal with 8 partners.