Region Selection based on Evidence Confidence for Localized Content-Based Image Retrieval
Wu-Jun Li,Dit-Yan Yeung
2009-01-01
Abstract:Over the past decade, multiple-instance learning (MIL) has been successfully utilized to model the localized content-based image retrieval (CBIR) problem, in which a bag corresponds to an image and an instance corresponds to a region in the image. However, existing feature representation schemes are not effective enough to describe the bags in MIL, which hinders the adaptation of sophisticated single-instance learning (SIL) methods for MIL problems. In this paper, we propose a conceptually simple but very powerful method for localized CBIR. First, a novel measure, called evidence confidence (EC), is proposed to select those regions which are most likely to support the labels of the images (i.e., bags). Then, based on the selected regions, a very effective feature representation scheme, which is also very computationally efficient and robust to labeling noise, is proposed to describe the bags. As a result, the MIL problem is converted into a standard SIL problem and traditional SIL methods, such as support vector machine (SVM), can be easily adapted for localized CBIR. Furthermore, using our feature representation scheme, a semi-supervised learning framework based on manifold regularization is proposed to incorporate unlabeled data into the training process. Experimental results on two challenging data sets show that our supervised learning method, called EC-SVM, outperforms the state-of-the-art methods and our semi-supervised learning framework, called EC-LapSVM, can lead to further performance improvement.