December 28, 2011
School of Engineering, Bengaluru
The Department of Telemedicine at the Amrita Institute of Medical Sciences sends out a trained technician to a remote area with an echo-cardiography machine. This way, even without the presence of a cardiologist on site, a large number of patients are easily screened for cardiovascular defects.
Now the efforts of S. Nandagopalan, Chairperson, Department of Computer Science, Bangalore Institute of Technology may make the screening even easier.
Nandagopalan, who is a registered doctoral student at Amrita Vishwa Vidyapeetham, is working on using bioinformatics to help reduce the incidence of heart diseases globally.
He is conducting his research under the guidance of Dr. T.S.B. Sudarshan, Chairperson, Department of Computer Science, Amrita School of Engineering, Bengaluru.
“Echocardiography is a popular technique for the diagnosis of heart abnormalities,” he shared.
“Generally, the images provide a wealth of clinically relevant and useful information, including the size and shape of the heart, its pumping capacity and the location and extent of any damage to its tissues. It is especially useful for assessing stenosis and regurgitation. However, current clinical practice requires manual intervention in both imaging and interpretation.”
“The ultra sound operator has to manually demarcate major anatomical structures like Left Ventricle, Right Ventricle, Left Atrium and Right Atrium and compute the length, diameter, area, functional shortening, stroke volume, ejection fraction, etc., after making the demarcations.”
“As the images are manually analyzed, a skilled and expert operator is needed to detect heart cavities accurately. Any error in the image quantification will lead to an incorrect diagnosis.”
Nandagopalan’s research focuses on using sophisticated computer vision and data mining based algorithms to enable automatic classification of cardiovascular ultrasound images.
Two research papers authored by him were recently published in the Journal of Computer Science and LNCS/LNAI.
Two others were accepted for publication in the International Journal of Data Mining and Bioinformatics and the Journal of Imaging Science.
“Our proposed method offers accurate and fast content-based archiving of relevant clinical results and support at ICU,” explained Nandagopalan.
“We are developing an implementable model that can analyze the echo image of a particular patient automatically and offer clinically relevant data for making appropriate decisions. For this, a series of steps have to be accomplished, starting from acquiring proper echo images of the patient and up to complex data mining and image processing tasks.”
“Once the echo images and/or video are available, they are preprocessed and segmented to obtain the Region of Interest. For this a number of efficient algorithms are being developed. One of the primary requirements of the current work is that all these algorithms be executed within the database environment as most patient data is already in the database. For this we are using SQL and PL/SQL procedures.”
“The outcome of this research should reduce the gap between the manual way of analysis and automated methodologies.”