Blood Smear Analysis

Analysis of Accuracy of Differential Count of White Blood Cells using Support Vector Machine

Mrs. Kiran J. Waghmare

Assistant Professor

Don Bosco College of Engineering, Fatorda, Goa. [e-mail] Gautam Gawas, Aishwarya Rana, Dilraj Velip , Tulsi Acharya, Vivek Kumar Jaiswal

BE Computer Engineering Students, Sem VIII

Don Bosco College of Engineering, Fatorda, Goa

ABSTRACT

Blood smear is a clinical test performed on microscopic digital images routinely investigated by hematologists to diagnose most blood diseases. Blood smear generally composed of Red blood cells (RBC), White blood cells (WBC) and Platelets. The quantities of WBC cell are counted in a sample blood smear and necessary information calculated manually by the

The count of 100 White blood cells with the help of microscope and then calculate the total occurrences of each type of white blood cells is called differential blood count.We ask that authors follow some simple guidelines. In essence, we ask you to make your paper look exactly like this document. The easiest way to do this is simply to download the template, and replace the content with your own material.

RELATED WORK

The analysis of the major issues in clinical labs is produced by Venkatalakshmi [8] .The number of RBC is very important to detect as well as to follow the treatment of many diseases like anaemia, leukemia etc.

The study of Image segmentation with different Segmentation methodare prposed by Rajwinder and Harpreet Kaur [7]. In this paper discussion about Generalized Hough Transform, Circulalar Hough Transform was done uusing knowledge based Learning.

Support Vector Machines introduced in COLT-92 by Boser, Guyon and Vapnik.Theorotically SVM is a well motivated algorithm to develop a Satistical Learning theory. SVM results in good performance in many applications like bioinformatics, text , image recognition, etc.Initially SVM popularized in the NIPS community, now an important and active field of all Machine Learning Research.[3] The special issues of Machine Learning Journal, and Journal of Mahine Learning

Proposed Model for WBC Differential count

Image Acquisition

Image Acquisition is the first step of the proposed model. Digitized images of the blood smear samples on the slides are acquired .A sets of 80 blood smear slide images were collected for this paper. A Image dataset was collected from Hospital in Goa, where normal peripheral blood slides under light microscope with 100× magnification; 80 blood smear images with white blood cells were captured by a Nikon D7200 high-definition color camera with a 24.0MP APS-c sensor and saved in JPG format of size 960 × 1,280 pixels at 15 pixels per 1 ?m resolution.

Image Preprocessing

Image pre-processing is significantly required to increase the reliability of an optical inspection at its lowest level of abstraction. These preprocessing operations do not increase the image information but it improves the image by removing distortions and noise. For example Normalization, Edge filters,, soft focus, user-specific filter, banalization, etc., Due to the use of several filter operations image can be intensify or reduce certain image details to enable some faster