Manifold learning for novelty detection and its application in gesture recognition

Abstract As a state-of-the-art novelty detection method, Kernel Null Foley–Sammon Transform (KNFST) could identify multiple known classes and detect novelties from an unknown class via single model. However, KNFST only captures the global information of training set. The local geometrical structure is neglected. In this paper, manifold incorporated into to solve issue. First, we use graphs depict for within-class scatter total scatter. Second, samples same are mapped point in null space projected directions (NPDs). proposed method can overcome weakness caused by ignoring class. experimental results on several toy benchmark datasets show that learning (MLND) superior KNFST.