Classification of image database using SVM with Gabor Magnitude

The CBIR term has been widely used to describe the process of retrieving desired images from a large collection on the basis of features (such as color, texture and shape) that can be extracted from the images themselves. In this paper, we have proposed an image retrieval system on the basis of classification using Support Vector Machine (SVM) which is implemented in MATLAB with the help of Gabor Filtered image features. In the proposed system, texture features are found by calculating the Standard Deviation of the Gabor Filtered image. A SVM classifier can be learned from training data of relevance images and irrelevance images marked by users. Using the classifier, the system can retrieve more images relevant to the query in the database efficiently. The proposed CBIR technique is implemented on a database having 1000 images spread across 11 categories and COIL image database having 1080 images spread across 15 categories. For each proposed CBIR technique 55 queries (5 per category) are fired on the database and net average precision and recall are computed. The results have shown performance improvement with higher precision and recall values, achieving crossover point as high as 89% with SVM technique as compared to image retrieval using Gabor Magnitude without SVM technique where the maximum crossover point is approximately 79%.