A Hierarchical Rule-Based Method for Image Segmentation using Maximum-Gradient Profiles
Alan C. F. Colchester, R. T. Ritchings, Nihal D. Kodikara · 1988
For each p i x e l in an image we determined the local direct ion of greatest grey l e v e l change and linked pixels in this direction to form maximum gradient prof i l e s (MGP). Adjacent MGP's were linked side-by-side t o f o r m e d g e sections. These edge s e c t i o n s corresponded very closely to those from a Canny operator. For the present application, which was concerned with detection and characterisation of blood vesse l s on angiograms, subsequent grouping was res tr ic ted to anti-symmetric edge sections which formed a ridge with high grey v a l u e s in the centre. The ridges corresponded closely with blood ves se l s shown in the image. The width of each ridge and the locus of i t s centre-line were d e r i v e d from a n t i-symmetric MGP p a i r s a n d corresponded very well with expert judgement of v e s s e l width and centre line location. Advantages of the MGP grouping method included explicit r epresenta t ion of the profiles of grey values normal to edges; ease of implementation of higher-order grouping stages which were natural extensions of lower level processes; and hierarchical data structure representing image features at a l l s ca l e s, allowing improved interaction of bottom-up and top-down processing.