A Medical Image Retrieval System using Gabor Barcode

JCHPS Special Issue 1: February 2017 www.jchps.com Page 212 A Medical Image Retrieval System using Gabor Barcode B. Nandhini*, S. Gowri Department of ECE, M. Kumarasamy College of Engineering, Karur, Tamil Nadu, India. *Corresponding author: E-Mail: [email protected] ABSTRACT Advances in digital imaging technologies and the increasing prevalence of picture archival systems have led to an exponential growth in the number of medical images generated and stored in hospitals during recent years. Thus, medical image retrieval plays a major role in medical diagnosis. The Digital images in the form of X-Rays, MRI, CT is probably one of the most important tools in medicine since it provides a method for diagnosis, monitoring drug treatment responses and disease management of patients with the advantage of being a very fast non-invasive procedure. It is essential develop effective and efficient medical image retrieval systems for diagnosis, research and educational purposes. In medical image database we could store X-Rays, MRI, CT images, type of disease and its diagnosis details of particular patient or even a small text comment concerning clinical relevant information. To retrieve this information query image can be applied for retrieving images from database. In Image retrieval systems, the feature in the image is extracted as feature vector. The separate training phase was carried out for database feature vector extraction. Recently, Radon codes are determined from the radon transform through local thresholding. In this paper, Gabor transform which is also a powerful tool for feature extraction of texture based information. This paper uses Gabor barcodes a new structure for the image annotation. Gabor barcodes was created from a query image with Gabor filters results in different barcode lengths. The image is retrieved from the database by final stage of Support Vector Machine (SVM) based image classification. Results obtained show that the proposed SVM classifier outperforms multilayer perceptron neural network.