Multitask Text-to-Visual Embedding with Titles and Clickthrough Data
This addresses the text-visual embedding problem in vision-language research, offering incremental improvements for retrieval applications.
The paper tackles the problem of text-visual embedding by proposing a method that uses both image titles and click-through data from search engines, along with a new triplet loss and hard negative sampling approach. The results show it outperforms existing methods and is effective for real-world text-to-visual retrieval.
Text-visual (or called semantic-visual) embedding is a central problem in vision-language research. It typically involves mapping of an image and a text description to a common feature space through a CNN image encoder and a RNN language encoder. In this paper, we propose a new method for learning text-visual embedding using both image titles and click-through data from an image search engine. We also propose a new triplet loss function by modeling positive awareness of the embedding, and introduce a novel mini-batch-based hard negative sampling approach for better data efficiency in the learning process. Experimental results show that our proposed method outperforms existing methods, and is also effective for real-world text-to-visual retrieval.