Detection of Leucocytes in Microscopic Images with Swarm Intelligence Algorithm

Author Name(s): M. Venkata Dasu, Dr.P.Subbaiah
Author Email: dassmarri@gmail.com

Abstract

Leukemia is a blood a cancer. It is also Known as Hematologic cancer. Leukemia is due to abnormal growth of blood cells by the bone marrow in human body. It normally affects the white blood cells (Leucocytes). This disease can be recognized and diagnosed by analyzing microscopic images. Generally blood cells are captured by microscopic images. Early detection of cancer increases the survival rate of the patient. To detect leukemia cells in blood smear hematologists finds difficulty and it is inaccurate and time consuming. So it requires fast and cost effective system to identify the disease. Image processing techniques are most useful for the analysis of microscopic images. These methods do not require special equipment for lab testing .Image processing techniques are fast and cheap. In this paper Image processing methods and optimization algorithms are used for the detection of leucocytes. ACO (Ant colony optimization) and PSO (Particle swarm optimization algorithm) are existing algorithms and a Cuckoo search Optimization algorithm is proposed method (CSO). By observing Statistical parameters the proposed method outperforms the existing methods. These results are obtained by simulating MATLAB software with I5 processor and 8GB RAM

Introduction

As of late, Digital image processing is connected to numerous applications which incorporate in remote sensing, biomedical and so on. Specifically, imaging applications are rising as another open door for modernization at the meeting point among computer science and medicine. Image processing is valuable to remove significant data from medicinal images and enhances the diagnosis of Leukemia in patients [1]. Malignant cells can be detected by human look and it takes lot of time and inaccurate also .Image processing plays important step in providing the correct form of diagnosis. Leukemia is the most dangerous disease, it can progress rapidly and if the patient not treated in time it may leads to fatal death within few months. Leukemia is mostly effects in children and adults. Generally majority malignancy cells start in body parts yet leukemia is the sort of disease which starts and develops in blood cells (1). Blood is essential substance without which metabolic elements of body is seriously influences. Bone marrow produces Blood cells. Blood cell comprises of three distinct kinds of cells such as WBC, RBC and Platelets. Erythrocytes (RBCs) are accountable for taking oxygen and CO2 away from tissues. There are three kinds of WBCs they are called as lymphocytes, monocytes, eosinophils, basophils and neutrophils [2-3]. Typically, leucocytes (WBCs) combat with the foreign bodies and keeps from contamination. Platelets help the coagulating of blood and controls hemorrhage (2) In Human system cell grows and multiply into new cells. In cancer disease, the old cells are not perished and remains in the blood. So the newly generated cells can’t have sufficient space to live and finally the functioning of blood distracts and generation of blood cells is irregular and uncontrolled [4]. In this work a novel method (Cuckoo Search algorithm) is proposed for the detection of cancerous (Malignant) cells accurately in microscopic medical images. Also in this work the obtained values from the proposed method are compared with the existing methods (ACO & PSO).

Conclusion

In this work, a heuristic optimization algorithm is adopted to identify the cancerous cells in microscopic blood smear. The preprocessing techniques in proposed model are median filter and CLAHE (Contrast Limited Adaptive Histogram Equalization)
method. The segmentation accuracy of the proposed method Cuckoo search algorithm is 97.58% is higher value than proposed methods. The advantage of proposed method is very accurate and less time consuming.

276 total views, 1 views today

Download PDF File

About the author: admin