期刊名称：International Journal of Innovative Research in Computer and Communication Engineering
印刷版ISSN：2320-9798
电子版ISSN：2320-9801
出版年度：2017
卷号：5
期号：2
页码：1708
DOI：10.15680/IJIRCCE.2017.0502087
出版社：S&S Publications
摘要：The framework for a system that automatically detects and classifies various diseases based on theimages generated in the medical industry using significant and biologically interpretable difference in features is beingproposed. The main objective of the proposed system is to automatically detect and classify various diseases based onthe images that has been inputted to the system which are mainly scanning, X-ray kind of reports. The system isexpected to predict what kind of a disease a person is more likely to be affected with based on his medical reportswhich are dominantly images. The image retrieval technique here is based on Content Based Image Retrieval (CBIR)using k-means clustering algorithm. The major steps include decomposition, feature extraction, clustering andprediction. Haar Wavelet Transform does the image decomposition and using mean and standard deviationcalculations feature vectors are constructed. Once query image is uploaded, similar processes are repeated and fromclusters formed, similarity calculations are performed. An additional feature to the system include Automatic LabelCorrection involving label correction in case of wrong uploads which otherwise could lead to wrong predictions andhence affect accuracy.
