Skip to Main Content
In this paper, we propose a simple yet efficient method for superpixel level object recognition on the bag-of-feature framework. Instead of using general classifiers for the superpixel categorization, we introduce local learning classifiers into our framework, so as to tackle the intraclass variation problem brought by superpixel based representations of objects. In addition, context information is used to make better performance by combining each superpixel with its most similar neighbors. We test our proposed method on Graz-02 datasets, and get results comparable to the state-of-the-art.