Morphological and textural analysis of centroblasts in low-thickness sliced tissue biopsies of follicular lymphoma.
Τίτλος | Morphological and textural analysis of centroblasts in low-thickness sliced tissue biopsies of follicular lymphoma. |
Publication Type | Journal Article |
Year of Publication | 2014 |
Authors | Michail, E., Dimitropoulos K., Koletsa T., Kostopoulos I., & Grammalidis N. |
Journal | Conf Proc IEEE Eng Med Biol Soc |
Volume | 2014 |
Pagination | 3374-7 |
Date Published | 2014 |
ISSN | 1557-170X |
Λέξεις κλειδιά | Algorithms, Biopsy, Cell Nucleolus, Cell Shape, Endothelial Cells, Humans, Image Processing, Computer-Assisted, Lymphoma, Follicular |
Abstract | This paper presents a new method for discriminating centroblast (CB) from non-centroblast cells in microscopic images acquired from tissue biopsies of follicular lymphoma. In the proposed method tissue sections are sliced at a low thickness level, around 1-1.5 μm, which provides a more detailed depiction of the nuclei and other textural information of cells usually not distinguishable in thicker specimens, such as 4-5 μm, that have been used in the past by other researchers. To identify CBs, a morphological and textural analysis is applied in order to extract various features related to their nuclei, nucleoli and cytoplasm. The generated feature vector is then used as input in a two-class SVM classifier with ε-Support Vector Regression and radial basis kernel function. Experimental results with an annotated dataset consisting of 300 images of centroblasts and non-centroblasts, derived from high-power field images of follicular lymphoma stained with Hematoxylin and Eosin, have shown the great potential of the proposed method with an average detection rate of 97.44%. |
DOI | 10.1109/EMBC.2014.6944346 |
Alternate Journal | Conf Proc IEEE Eng Med Biol Soc |
PubMed ID | 25570714 |