Scene and Object Recognition with Supervised Nonlinear Neighborhood Embedding

Image category recognition is important to access visual information on the level of objects and scene types. In this paper, we develop a Supervised Nonlinear Neighborhood Embedding (SNNE) subspace algorithm of different visual features for object and scene recognition, which learns an adaptive nonlinear subspace by preserving the neighborhood structure of the visual feature space. In the proposed subspace algorithm, we combine the idea of nonlinear kernel mapping and preserving the neighborhood structure of the samples, so it can not only gain a perfect approximation of the nonlinear image manifold, but also enhance within-class neighborhood information. So, the proposed SNNE algorithm models the ensemble of visual features to a more discriminative space for category recognition, and at the same time, can effectively combine several visual features to improve recognition rate. The proposed method is evaluated by using the scene database (SIMPLicity) and object recognition database (Caltech). We confirm that the proposed method is much better than state-of-the-art methods only with simple visual features.