Hardware approach to distance transform

The computation of Euclidean distances, as opposed to digital distances (4-distance and 8-distance) involves a large amount of computations. Significant reduction in complexity can be achieved by plotting circles with increasing diameters around a pixel and examining the pixels on the circumferences. Two hardware configurations for computing the Euclidean distances using circles are proposed. The first configuration processes the bitmap serially. The circumferences are examined in parallel. The distances are found after one scan of the bitmap. The second is a cellular configuration and processes the bitmap in parallel. Pixels on the circumferences are examined serially.