AN EFFECTIVE MACHINE LEARNING ALGORITHM FOR TEXTURE BASED MEDICAL IMAGE RETRIEVAL SYSTEM

Abstract

In the present digital world, an image databases are increasing enormously across the world. An effective image retrieval approach is needed for
utilizing this massive databases. An extensive research efforts have been conducted in the field of Content-Based Medical Image
Retrieval(CBMIR) system. This paper analysed a novel evolutionary approach to extract texture features for CBMIR application.The selected
texture features are Local Octal pattern(LOP)in which extracted features are formed as feature vector database. A machine learning algorithms
are analysed for feature selection and classification problems. To reduce the high dimensional texture features, Grey Wolf Optimization(GWO)
is used to select the best features. A classification algorithm is used as an Evaluation Criteria, for identifying the best subset of features. Fuzzy based
Relevance Vector Machine(FRVM) based classification algorithm is applied to classify the subset of texture features of the images. Euclidean
Distance(ED) is used as similarity measurement techniques, to identify the similarity between the query image and the classified image feature
database.To evaluate the retrieval performance, an experiments have been conducted on medical image dataset. The Precision and Recall is used
as a performance metrics to evaluate the CBMIR systems.