Automatic recognition of myeloma cells in microscopic images using bottleneck algorithm, modified watershed and SVM classifier
- 19 January 2016
- journal article
- Published by Wiley in Journal of Microscopy
- Vol. 261 (1) , 46-56
Abstract
Summary: Plasma cells are developed from B lymphocytes, a type of white blood cells that is generated in the bone marrow. The plasma cells produce antibodies to fight with bacteria and viruses and stop infection and disease. Multiple myeloma is a cancer of plasma cells that collections of abnormal plasma cells (myeloma cells) accumulate in the bone marrow. The definitive diagnosis of multiple myeloma is done by searching for myeloma cells in the bone marrow slides through a microscope. Diagnosis of myeloma cells from bone marrow smears is a subjective and time‐consuming task for pathologists. Also, because of depending on final decision on human eye and opinion, error risk in decision may occur. Sometimes, existence of infection in body causes plasma cell's increment which could be diagnosed wrongly as multiple myeloma. The computer diagnostic process will reduce the diagnostic time and also can be worked as a second opinion for pathologists.This study presents a computer‐aided diagnostic method for myeloma cells diagnosis from bone marrow smears. At first, white blood cells consist of plasma cells and other marrow cells are separated from the red blood cells and background. Then, plasma cells are detected from other marrow cells by feature extraction and series of decision rules. Finally, normal plasma cells and myeloma cells could be classified easily by a classifier.This algorithm is applied on 50 digital images that are provided from bone marrow aspiration smears. These images contain 678 cells: 132 normal plasma cells, 256 myeloma cells and 290 other types of marrow cells. Applying the computer‐aided diagnostic method for identifying myeloma cells on provided database showed a sensitivity of 96.52%; specificity of 93.04% and precision of 95.28%.Keywords
This publication has 14 references indexed in Scilit:
- Extraction of Nucleolus Candidate Zone in White Blood Cells of Peripheral Blood Smear Images Using Curvelet TransformComputational and Mathematical Methods in Medicine, 2012
- Clump splitting via bottleneck detection and shape classificationPattern Recognition, 2012
- Automatic white blood cell segmentation using stepwise merging rules and gradient vector flow snakeMicron, 2011
- Automatic thresholding for defect detectionPattern Recognition Letters, 2006
- Plasma Cells2006
- Model-based segmentation of nucleiPattern Recognition, 2000
- BINARY IMAGE SEGMENTATION OF AGGREGATES BASED ON POLYGONAL APPROXIMATION AND CLASSIFICATION OF CONCAVITIESPattern Recognition, 1998
- Recursive splitting of active contours in multipleclump segmentationElectronics Letters, 1996
- Clump splitting through concavity analysisPattern Recognition Letters, 1994
- Morphological segmentationJournal of Visual Communication and Image Representation, 1990