Voting conditional random fields for multi-label image classification

In our real world, there usually exist several different objects in one image, which brings intractable challenges to the traditional pattern recognition methods to classify the images. In this paper, we introduce a Conditional Random Fields (CRFs) model to deal with the Multi-label Image Classification problem. Considering the correlations of the objects, a second-order CRFs is constructed to capture the semantic associations between labels. Different initial feature weights are set to introduce the voting techniques for a better performance. We evaluate our methods on MSRC dataset and demonstrate high precision, recall and F₁ measure, showing that our method is competitive.