2D and 3D active shape model with SURF algorithm for OBJECT retrieval: Content Based Image Retrieval

Abstract

The Content Based Image Retrieval plays an important role in image retrieval methods. In this CBIR method mostly uses the different type of approaches, where the active shape model is mostly used by these image retrieval methods. In this paper, we present an approach that simultaneously uses the Active Shape Model in 3D environment method with SURF (Speeded Up Robust Features) algorithm for improving the performance of the image retrieval system. In this active shape method we apply the 3D model to 2D image for extracting some parts from the image. Then we add the SURF algorithm to this 3D active shape model for detect and describe local features in images, Our retrieval method provides both good retrieval performance and greater robustness compared to 3D Model and Part Information For Improving Content based Vehicle Retrieval, In Plane Rotation and Scale Invariant Clustering methods, Wavelet Transform coefficients distribution in content based image retrieval methods. Retrieval efficiency is given for different databases including a car database, a MR images and a face database. Finally the retrieval efficiency is higher than 90% for some cases using an optimization process.