Digital blood image processing and fuzzy clustering for detection and classification of atypical lymphoid B cells
Document typeConference report
Rights accessOpen Access
Automated systems for digital peripheral blood (PB) cell analysis operate most effectively in non-pathological samples. The paper deals with the automatic classification of atypical lymphoid cells using digital image processing. The problem has been approached through a 3-step procedure: 1) Watershed segmentation of nucleus, cytoplasm and peripheral cell zone; 2) feature extraction for each region; and 3) classification using fuzzy c-means. The paper has proposed a new methodology that has been able to automatically classify with high precision three types of lymphoid cells: normal, Hairy Cell Leukemia cells and Chronic Lymphocytic Leukemia cells. This methodology, combining human medical expertise with mathematical and engineering tools, may contribute to improve the efficiency of the hematology laboratory.
CitationAlférez Baquero, Edwin Santiago [et al.]. Digital blood image processing and fuzzy clustering for detection and classification of atypical lymphoid B cells. A: "Jornades de recerca EUETIB". Barcelona: 2013, p. 1-12.