A Case Study: Circles and Ellipses

One of the most striking things concerning the plethora of literature relating to computationally realistic implementations of the Hough Transform is that the majority of effort is directed towards straight line detection. Such implementations can seldom be generalized to include the detection of shapes having more than two defining parameters. Moreover, the detection of more complicated shapes is rarely dealt with other than by the use of software solutions. The Probabilistic Houghs offer a unique opportunity for generalization to n dimensions. In particular, the Dynamic Generalized Hough Transform has an algorithm that is uniquely suited to parallel implementation irrespective of the number of parameters under detection.