Bilkent University
Department of Computer Engineering


Unsupervised segmentation of colon glands using subgraphs


Orhun Alp Oral
MSc Student
Computer Engineering Department
Bilkent University

Analysis of biopsies which contain glandular structures, segmentation of glands is an important step. However, it is a challenging problem as the variation in staining, fixation, and sectioning procedures lead to a considerable amount of artifacts and variances in tissue sections, which may result in huge variances in gland appearances. In this work, we report a new approach for gland segmentation. This approach has two main contributions. First, it introduces a new way of defining disconnected subgraphs to model the tissue images. Rather than on image pixels, nodes of these subgraphs are defined from tissue components. These components are approximately represented by tissue objects, and expected to be less vulnerable to noise and variations that are typically observed at the pixel level of tissue images.Second, as distinct from the previous methods, which are partitions the tissue graph into subgraphs to segment. Segmentation algorithm in this work proposes to obtain each subgraph segmentation by using a gaussian classifier function among from the image pixels.


DATE: 30 December, 2013, Monday @ 16:50