Output details

Maximum likelihood estimation of vessel parameters from scale space analysis

<23> This paper describes a novel solution to retinal vessel segmentation and the work resulted from PPARC funding (PPA/1/S/2002/00659:C). A model based approach for both vessels and background is proposed that is shown to be successful on normal retinal images in addition to those where features in the background (such as choroid vessels) provide the image processing challenge. My previously published work on the analysis of retinal images containing choroid vessels aided the development of this approach. My contribution was as co-investigator on the original PPARC award and as co-supervisor of the post-doctoral researchers who implemented the algorithms (Ng & Clay).