A Proposed Novel Method for Detection and Classification of Leukemia using Blood Microscopic Images

• Published in: International journal of advanced research in computer science 2017-03, Vol.8 (3)
• Main Authors: Begum, A R Jasmine, T Abdul Razak
• Format: Article
• Language: English
• Subjects: Blood; Cancer; Clusters; Erythrocytes; Feature extraction; Image classification; Image detection; Image enhancement; Image quality; Image reconstruction; Image segmentation; Leukemia; Leukocytes; Mean square errors; Nuclei (cytology); Pattern recognition; Signal to noise ratio
• ISSN: 0976-5697

Leukemia is a form of cancer in the blood cells. Most forms of Leukemia occur in the White Blood Cells. The aim of the study is to recognize leukemia White Blood Cells from blood microscopic images through segmentation. The proposed work, an Enhanced Hybrid Fuzzy C-Means with Cluster Center Estimation is to separate the nucleus from the White Blood Cell image which is received as an output of the Hybrid Fuzzy C-Means with Cluster Center Estimation algorithm. The structural image modelling technique called morphological operation is utilized to extract the nucleus followed by feature extraction to extract the geometric features of the nuclei. The Support Vector Machine which is an empirically good performance pattern recognition tool to perform the classification and prediction of cancer affected blood cell images. An analysis is carried by considering the extracted feature of White Blood Cell and nucleus. Finally the performance of this method is evaluated on the basis of image quality measures such as Peak Signal to Noise Ratio and Mean Square Error. Thus, this method results in a high Peak Signal to Noise Ratio and low Mean Square Error, which indicates the better reconstruction of the images in comparison with other existing methods.