TY - GEN
T1 - Visual Grounding in Remote Sensing Images
AU - Sun, Yuxi
AU - Feng, Shanshan
AU - Li, Xutao
AU - Ye, Yunming
AU - Kang, Jian
AU - Huang, Xu
N1 - Publisher Copyright:
© 2022 ACM.
PY - 2022/10/10
Y1 - 2022/10/10
N2 - Ground object retrieval from a large-scale remote sensing image is very important for lots of applications. We present a novel problem of visual grounding in remote sensing images. Visual grounding aims to locate the particular objects (in the form of the bounding box or segmentation mask) in an image by a natural language expression. The task already exists in the computer vision community. However, existing benchmark datasets and methods mainly focus on natural images rather than remote sensing images. Compared with natural images, remote sensing images contain large-scale scenes and the geographical spatial information of ground objects (e.g., longitude, latitude). The existing method cannot deal with these challenges. In this paper, we collect a new visual grounding dataset, called RSVG, and design a new method, namely GeoVG. In particular, the proposed method consists of a language encoder, image encoder, and fusion module. The language encoder is used to learn numerical geospatial relations and represent a complex expression as a geospatial relation graph. The image encoder is applied to learn large-scale remote sensing scenes with adaptive region attention. The fusion module is used to fuse the text and image feature for visual grounding. We evaluate the proposed method by comparing it to the state-of-the-art methods on RSVG. Experiments show that our method outperforms the previous methods on the proposed datasets. https://sunyuxi.github.io/publication/GeoVG
AB - Ground object retrieval from a large-scale remote sensing image is very important for lots of applications. We present a novel problem of visual grounding in remote sensing images. Visual grounding aims to locate the particular objects (in the form of the bounding box or segmentation mask) in an image by a natural language expression. The task already exists in the computer vision community. However, existing benchmark datasets and methods mainly focus on natural images rather than remote sensing images. Compared with natural images, remote sensing images contain large-scale scenes and the geographical spatial information of ground objects (e.g., longitude, latitude). The existing method cannot deal with these challenges. In this paper, we collect a new visual grounding dataset, called RSVG, and design a new method, namely GeoVG. In particular, the proposed method consists of a language encoder, image encoder, and fusion module. The language encoder is used to learn numerical geospatial relations and represent a complex expression as a geospatial relation graph. The image encoder is applied to learn large-scale remote sensing scenes with adaptive region attention. The fusion module is used to fuse the text and image feature for visual grounding. We evaluate the proposed method by comparing it to the state-of-the-art methods on RSVG. Experiments show that our method outperforms the previous methods on the proposed datasets. https://sunyuxi.github.io/publication/GeoVG
KW - dataset
KW - object retrieval
KW - referring expression
KW - remote sensing
KW - visual grounding
UR - https://www.scopus.com/pages/publications/85151115028
U2 - 10.1145/3503161.3548316
DO - 10.1145/3503161.3548316
M3 - 会议稿件
AN - SCOPUS:85151115028
T3 - MM 2022 - Proceedings of the 30th ACM International Conference on Multimedia
SP - 404
EP - 412
BT - MM 2022 - Proceedings of the 30th ACM International Conference on Multimedia
PB - Association for Computing Machinery, Inc
T2 - 30th ACM International Conference on Multimedia, MM 2022
Y2 - 10 October 2022 through 14 October 2022
ER -