Authors: N.R. Shanker; A. Ezhil; S. Archana
Addresses: Sree Sastha Institute of Engineering and Technology, Sree Sastha Nagar, Chennai – Bangalore Highway, Chembarambakkam, Chennai – 600123, India. ' Vel Tech High Tech Dr. Rangarajan Dr. Sakunthala Engineering College, #60, Avadi – Alamathi Road, Vellanur, Avadi, Chennai – 600 062, India. ' Vel Tech High Tech Dr. Rangarajan Dr. Sakunthala Engineering College, #60, Avadi – Alamathi Road, Vellanur, Avadi, Chennai – 600 062, India
Abstract: A novel identification and detection of cholesterol in the human body by non-invasive method using image processing is presented in this paper. Different sample of images with and without cholesterol are taken for the study. These images are analysed using mean algorithms in image processing to detect the cholesterol levels. The image of the patients' finger region is taken as sample images, along with their laboratory tested values of cholesterol. A database of different range of cholesterol values is created using these images. The sample images of different age groups are collected for the purpose of easy image analysis and accuracy. In image processing, the image analysis is done in various methods such as mean algorithm, median, standard deviation, histogram analysis, grey slicing method, etc. It was found that the mean algorithm is suitable for the non-invasive method of detecting the cholesterol levels. The mean value of the test image is then compared to the mean value of the images in the database to determine the cholesterol value. From the results it is found that the cholesterol mean values are proportional to the laboratory values. Thereby the correlations table is formulated.
Keywords: image processing; cholesterol detection; correlation; average mean algorithms; non-invasive methods; cholesterol values.
International Journal of Medical Engineering and Informatics, 2012 Vol.4 No.3, pp.223 - 230
Published online: 02 Aug 2012 *Full-text access for editors Access for subscribers Purchase this article Comment on this article