Guy L. Scott

Learn More
We describe a widely applicable method of grouping-or clustering-image features (such as points, lines, corners, flow vectors and the like). It takes as input a "proximity matrix" H-a square, symmetric matrix of dimension N (where N is the number of features). The element i,j of H is an initial estimate of the "proximity" between the i-th and yth features.(More)
A relaxation algorithm for the computation of optic flow at edge elements (edgels) is presented. Flow is estimated only at intensity edges of the image. Edge elements , extracted from an intensity image, are used as the basis for the algorithm. A matching strength or weight surface is computed around each edgel and neighbourhood support obtained to enhance(More)
Hildreth described a method of extracting the normal component of optic flow along a moving closed contour. She also described an algorithm-based on gradient descent of recovering the smoothest full flow field consistent with the normal flow. Her iterative method is somewhat slow. It also leaves open the question of what, if anything, we are going to say(More)
A highly efficient means of describing curves, surfaces or other configurations in space (or space-time) is to express the position vector as a sum of functions defined over some some interval in a space of one or more "im-plicit variables". The most familiar forms of these basis functions are the polynomial, trigonometric and super-quadric. There seems no(More)
This paper describes a simple, fast and robust method of obtaining a depth field from a large number of (closely spaced) views of a scene. The method-"synthetic aperture depth-from-focus"-involves superimposing images in such a way as to obtain a synthetic "blurred" image similar to that which would be obtained by a single camera with a very large aperture(More)
  • 1