CBIR by cascading features & SVM

This paper investigates different methods of representing shape and texture in content-based image retrieval. We have combined five features set in our work and these are trained and classified with SVM (support vector machine) classifier which makes use of machine learning technology. We combined histogram features, texture features (GLCM features), wavelet features, Gabor features, and statistical features, which makes use of global and local features. A database of 1000 images (Wang database) of 10 different classes is used to extract all features vector for each image and stored in our database so that SVM can use it to classify the query image. By using these features set, we are able to reach up to 97.53% classification accuracy.