Effective Detection of Nucleus in Cytodiagnosis by Employing Combinatorial Fuzzy Hough Transform

Hough transform is an effective method for line detection in a digital image. However, the conventional Hough transform can not be applied well enough for an image corrupted by heavy noise. This paper describes a combinatorial fuzzy Hough transform, which enables a detection of not only a blurred line, but also a segment of a line in a very noisy image. It is then extended in order that it can further detect a blurred circle. The effectiveness of the present method has been confirmed by applying it to an actual microphotograph for cytodiagnosis. It can detect the nuclei in a cell cluster in a very noisy medical image.