Dual Attention on Pyramid Feature Maps for Image Captioning
This work addresses the problem of generating natural sentences from images for visual-semantic understanding in multimedia, representing an incremental improvement through modular attention mechanisms.
The paper tackles image captioning by applying dual attention on pyramid feature maps to better localize relevant regions and recalibrate feature importance, achieving impressive results on Flickr8K, Flickr30K, and MS COCO datasets with promising performance in single-model mode.
Generating natural sentences from images is a fundamental learning task for visual-semantic understanding in multimedia. In this paper, we propose to apply dual attention on pyramid image feature maps to fully explore the visual-semantic correlations and improve the quality of generated sentences. Specifically, with the full consideration of the contextual information provided by the hidden state of the RNN controller, the pyramid attention can better localize the visually indicative and semantically consistent regions in images. On the other hand, the contextual information can help re-calibrate the importance of feature components by learning the channel-wise dependencies, to improve the discriminative power of visual features for better content description. We conducted comprehensive experiments on three well-known datasets: Flickr8K, Flickr30K and MS COCO, which achieved impressive results in generating descriptive and smooth natural sentences from images. Using either convolution visual features or more informative bottom-up attention features, our composite captioning model achieves very promising performance in a single-model mode. The proposed pyramid attention and dual attention methods are highly modular, which can be inserted into various image captioning modules to further improve the performance.