Color clustering segmentation framework for image analysis of malignant lymphoid cells in peripheral blood.
Current computerized image systems are able to recognize normal blood cells in peripheral blood, but fail with abnormal cells like the classes of lymphocytes associated to lymphomas. The main challenge lies in the subtle differences in morphologic characteristics among these classes, which requires...
| Publicado en: | Medical & Biological Engineering & Computing Vol. 57; no. 6; pp. 1265 - 1284 |
|---|---|
| Autores principales: | , , , , |
| Formato: | Journal Article |
| Publicado: |
Springer Nature
Jun2019
|
| Acceso en línea: | Ver este registro en EBSCOhost |
| fields | @attributes: recordID: 1 pdfLink: plink: https://search.ebscohost.com/login.aspx?direct=true&db=ccm&AN=136505526&site=ehost-live header: @attributes: shortDbName: ccm uiTerm: 136505526 longDbName: CINAHL Complete uiTag: AN controlInfo: bkinfo: dissinfo: jinfo: jid: 01400118 PO0 jtl: Medical & Biological Engineering & Computing issn: 01400118 maglogo: N pubinfo: dt: Jun2019 vid: 57 iid: 6 pid: 237 pub: Springer Nature place: New York, New York artinfo: ui: 136505526 136505526 NLM30730028 10.1007/s11517-019-01954-7 NLM30730028 136505526 ppf: 1265 ppct: 19 formats: fmt: – @attributes: type: T – @attributes: type: P tig: atl: Color clustering segmentation framework for image analysis of malignant lymphoid cells in peripheral blood. aug: au: Alférez, Santiago Merino, Anna Acevedo, Andrea Puigví, Laura Rodellar, José affil: Department of Mathematics, EEBE, Technical University of Catalonia, Street Eduard Maristany 6-12, 08019, Barcelona, Spain sug: subj: Algorithms Lymphoma Blood Leukemia Blood Image Processing, Computer Assisted Lymphocytes Pathology Lymphoma Pathology Color Automation Leukemia Pathology Cell Nucleus Pathology Cluster Analysis Ways of Coping Questionnaire ab: Current computerized image systems are able to recognize normal blood cells in peripheral blood, but fail with abnormal cells like the classes of lymphocytes associated to lymphomas. The main challenge lies in the subtle differences in morphologic characteristics among these classes, which requires a refined segmentation. A new efficient segmentation framework has been developed, which uses the image color information through fuzzy clustering of different color components and the application of the watershed transformation with markers. The final result is the separation of three regions of interest: nucleus, entire cell, and peripheral zone around the cell. Segmentation of this zone is crucial to extract a new feature to identify cells with hair-like projections. The segmentation is validated, using a database of 4758 cell images with normal, reactive lymphocytes and five types of malignant lymphoid cells from blood smears of 105 patients, in two ways: (1) the efficiency in the accurate separation of the regions of interest, which is 92.24%, and (2) the accuracy of a classification system implemented over the segmented cells, which is 91.54%. In conclusion, the proposed segmentation framework is suitable to distinguish among abnormal blood cells with subtile color and spatial similarities. Graphical Abstract The segmentation framework uses the image color information through fuzzy clustering of different color components and the application of the watershed transformation with markers (Top). The final result is the separation of three regions of interest: nucleus, entire cell, and peripheral zone around the cell. The procedure is also validated by the implementation of a system to automatically classify different types of abnormal blood cells (Bottom). pubtype: Academic Journal doctype: Journal Article ougenre: Article language: English refInfo: holdings: @attributes: islocal: N |
|---|