Segmentation of cell nuclei in tissue by combining seeded watersheds with gradient information
2003 (English)In: Proceedings of SCIA-03: Scandinavian Conference on Image Analysis, 2003, 408-414 p.Conference paper (Refereed)
This paper deals with the segmentation of cell nuclei in tissue. We present a region-based segmentation method where seeds representing object- and background-pixels are created by morphological filtering. The seeds are then used as a starting point for watershed segmentation of the gradient magnitude of the original image. Over-segmented objects are thereafter merged based on the gradient magnitude between the adjacent objects. The method was tested on a total of 726 cell nuclei in 7 images, and 95% correct segmentation was achieved.
Place, publisher, year, edition, pages
2003. 408-414 p.
Computer Vision and Robotics (Autonomous Systems)
IdentifiersURN: urn:nbn:se:uu:diva-46181OAI: oai:DiVA.org:uu-46181DiVA: diva2:74089