Visual Object Localization in Image Collections

Abstract

The research of object localization is active in the field of visual object category. In this paper, we focus on object localization in a given special category dataset. We propose to exploit the context aware category discovery for object localization without any labeled examples. Firstly, the image is segmented based on a multiple segmentation algorithm. Secondly, these generated regions are clustered by spectral clustering method to find the category pattern based on the context of the dataset and the saliency. Thirdly, the object is localized based on the weakly supervised learning algorithm. To justify the effectiveness of the proposed method, the detection precision is employed to evaluate the performance of our approach. The experimental results demonstrate that our approach is promising in object localization with unsupervised learning method.