CBIR involves searching for similar images for a query image in an image database. The proposed\nmethodology aims at improving the classification and retrieval accuracy of images. Wavelet\nHistograms (WH) is used to design a simple and efficient CBIR system with good performance\nwithout using any intensive image processing feature extraction technique. The unique indexed\ncolor histogram and wavelet decomposition based horizontal, vertical and diagonal image\nattributes serve as the main features for the retrieval system. In this work, those distinct image\nfeatures are used for classification and retrieval using support vector machine (SVM).The\nperformance of the proposed content based image classification and retrieval system is evaluated\nwith the standard SIMPLIcity dataset. The performance of the system is measured with precision\nas the metric since the other compared methods use it as the main metric. For validating results,\nholdout validation and k-fold cross validation are used. The proposed system performs obviously\nbetter than SIMPLIcity and all the other compared methods.

Login to Download PDF

Please login with your Paper ID and password to access the full PDF.

🔑 Login to Download