摘要
The results of automatic classification of the peripheral blood leukocytecategories are presented.This classification was performed by means of digital imageprocessing and pattern recognition techniques.In this paper,an automatic microscopecellular analysis system for this purpose is given and a set of algorithms for leukocyterecognitions,which includes individual cell’s detection,scene segmentation,featureextraction,selection and classifier design,is proposed.As a result of test on 3000leukoeytes digitized directly from 21 peripheral blood smears,92 percent of themwere classified correctly in this system.The categories were band neutrophils,seg-mented neutrophils,monocytes,eosinophils,basophils and lyrnphoeytes.
The results of automatic classification of the peripheral blood leukocyte categories are presented.This classification was performed by means of digital image processing and pattern recognition techniques.In this paper,an automatic microscope cellular analysis system for this purpose is given and a set of algorithms for leukocyte recognitions,which includes individual cell's detection,scene segmentation,feature extraction,selection and classifier design,is proposed.As a result of test on 3000 leukoeytes digitized directly from 21 peripheral blood smears,92 percent of them were classified correctly in this system.The categories were band neutrophils,seg- mented neutrophils,monocytes,eosinophils,basophils and lyrnphoeytes.
