Paper Title
Multi-Tiered Approach For Content Based Image Retrieval Using SVM Classifier

Abstract
Abstract -A multi-tiered content-based image retrieval system for images using a reference database containing images of more than one category (flowers and vehicles) is considered. The multitier approach classifies and retrieves images involving their specific subtypes, which are mostly difficult to discriminate and classify. Initially, color and texture features are extracted and a feature vector is formed. The color features are extracted using color moments. The texture features are extracted using GLCM. The feature vector consists of both color and texture features. In the proposed system, the first tier uses SVM classifier to classify the category of images. Then SVM classifier is followed by KNN (K-nearest neighbor) which searches the corresponding database, index will be computed by similarity feature matching. The query image is classified by the classifier to a particular class and the relevant images are retrieved from the database. This system retrieves the most similar types in the image level by enabling multi-image queries in order to ensure the consistency among the retrieved images. The system enables multi-image query in order to protect the semantic consistency among the retrieved images.