Towards local visual modeling for image captioning

In this paper, we study the local visual modeling with grid features for image captioning, which is critical for generating accurate and detailed captions. To achieve this target, we propose a Locality-Sensitive Trans-former Network (LSTNet) with two novel designs, namely Locality-Sensitive Attention (LSA) and Locality-Sensitive Fusion (LSF). LSA is deployed for the intra-layer interaction in Transformer via modeling the relationship between each grid and its neighbors. It reduces the difficulty of local object recognition during captioning. LSF is used for inter-layer information fusion, which aggregates the information of different encoder layers for cross-layer semantical complementarity. With these two novel designs, the proposed LSTNet can model the local visual information of grid features to improve the captioning qual-ity. To validate LSTNet, we conduct extensive experiments on the competitive MS-COCO benchmark. The experimental results show that LSTNet is not only capable of local visual modeling, but also outperforms a bunch of state-of-the-art captioning models on offline and online testings, i.e., 134.8 CIDEr and 136.3 CIDEr, respectively. Besides, the generalization of LSTNet is also verified on the Flickr8k and Flickr30k datasets. The source code is available on GitHub: https://www.github.com/xmu-xiaoma666/LSTNet .(c) 2023 Elsevier Ltd. All rights reserved.