Integrating Scene Text and Visual Appearance for Fine-Grained Image Classification

Text in natural images contains rich semantics that is often highly relevant to objects or scene. In this paper, we focus on the problem of fully exploiting scene text for visual understanding. The main idea is combining word representations and deep visual features in a globally trainable deep convolutional neural network. First, the recognized words are obtained by a scene text reading system. Next, we combine the word embedding of the recognized words and the deep visual features into a single representation that is optimized by a convolutional neural network for fine-grained image classification. In our framework, the attention mechanism is adopted to compute the relevance between each recognized word and the given image, which further enhances the recognition performance. We have performed experiments on two datasets: con-text dataset and drink bottle dataset, which are proposed for fine-grained classification of business places and drink bottles, respectively. The experimental results consistently demonstrate that the proposed method of combining textual and visual cues significantly outperforms classification with only visual representation. Moreover, we have shown that the learned representation improves the retrieval performance on the drink bottle images by a large margin, making it potentially powerful in product search.