Qualitative and Quantitative Investigation of Bio Tissues using Microscopy and Data Mining

Preview this chapter:

Qualitative and Quantitative Investigation of Bio Tissues using Microscopy and Data Mining, Page 1 of 1

< Previous page | Next page > /docserver/preview/fulltext/9789815036930/chapter-3-1.gif

<div&gt;The effects of glucose and salt on white blood cells, red blood cells, and&lt;/div&gt;<div&gt;platelets (PLTs) in the blood of a leukemic patient by using a white light microscope&lt;/div&gt;<div&gt;have been investigated for different concentrations (0 mM to 500 mM) of glucose and&lt;/div&gt;<div&gt;salt. It has been revealed that the shape of erythrocytes, leukocytes, and platelets&lt;/div&gt;<div&gt;changes and forms aggregates. Increasing the concentration of sodium chloride causes&lt;/div&gt;<div&gt;an increase in the rouleaux formation and aggregation of platelets. The comparison of&lt;/div&gt;<div&gt;CBC reports of these samples with and without analytes shows that total leukocyte&lt;/div&gt;<div&gt;count (TLC) decreases gradually towards normal ranges of leukocytes, which is&lt;/div&gt;<div&gt;favorable in the treatment of leukemia; at the same time, decreased level of hemoglobin&lt;/div&gt;<div&gt;HGB, mean corpuscular hemoglobin (MCH), mean corpuscular hemoglobin&lt;/div&gt;<div&gt;concentration (MCHC) and increased level of red blood cell (RBCs) causes a reduction&lt;/div&gt;<div&gt;in the oxygen supply, which is in favor of cancer growth and anemia.<br&gt;<br&gt;</div&gt;<div&gt;In the second set of work, a computer-aided system was planned for automatic&lt;/div&gt;<div&gt;classification of ultrasound kidney diseases and ultrasound liver (i.e., cirrhosis). Two&lt;/div&gt;<div&gt;types of images were considered normal and chronic. By using the data mining&lt;/div&gt;<div&gt;technique, the statistical features were extracted to differentiate between normal and&lt;/div&gt;<div&gt;abnormal ultrasonic kidney images. By using feature extraction software (FES), a set of&lt;/div&gt;<div&gt;statistical features were extracted from the region of interest of each image at different&lt;/div&gt;<div&gt;frame rates. The data sets which were obtained using FES at different frame rates were&lt;/div&gt;<div&gt;then classified by using Weka. These extracted feature results were classified by using&lt;/div&gt;<div&gt;Weka and a 96.5% correct classification rate was obtained. The difference between the&lt;/div&gt;<div&gt;values of these features was useful to identify between normal and abnormal images.</div&gt;