TY - JOUR
T1 - STAR
T2 - A First-Ever Dataset and a Large-Scale Benchmark for Scene Graph Generation in Large-Size Satellite Imagery
AU - Li, Yansheng
AU - Wang, Linlin
AU - Wang, Tingzhu
AU - Yang, Xue
AU - Luo, Junwei
AU - Wang, Qi
AU - Deng, Youming
AU - Wang, Wenbin
AU - Sun, Xian
AU - Li, Haifeng
AU - Dang, Bo
AU - Zhang, Yongjun
AU - Yu, Yi
AU - Yan, Junchi
N1 - Publisher Copyright:
© 1979-2012 IEEE.
PY - 2025
Y1 - 2025
N2 - Scene graph generation (SGG) in satellite imagery (SAI) benefits promoting understanding of geospatial scenarios from perception to cognition. In SAI, objects exhibit great variations in scales and aspect ratios, and there exist rich relationships between objects (even between spatially disjoint objects), which makes it attractive to holistically conduct SGG in large-size very-high-resolution (VHR) SAI. However, there lack such SGG datasets. Due to the complexity of large-size SAI, mining triplets < subject, relationship, object > heavily relies on long-range contextual reasoning. Consequently, SGG models designed for small-size natural imagery are not directly applicable to large-size SAI. This paper constructs a large-scale dataset for SGG in large-size VHR SAI with image sizes ranging from 512 × 768 to 27 860 × 31 096 pixels, named STAR (Scene graph generaTion in lArge-size satellite imageRy), encompassing over 210K objects and over 400K triplets. To realize SGG in large-size SAI, we propose a context-aware cascade cognition (CAC) framework to understand SAI regarding object detection (OBD), pair pruning and relationship prediction for SGG. We also release a SAI-oriented SGG toolkit with about 30 OBD and 10 SGG methods which need further adaptation by our devised modules on our challenging STAR dataset.
AB - Scene graph generation (SGG) in satellite imagery (SAI) benefits promoting understanding of geospatial scenarios from perception to cognition. In SAI, objects exhibit great variations in scales and aspect ratios, and there exist rich relationships between objects (even between spatially disjoint objects), which makes it attractive to holistically conduct SGG in large-size very-high-resolution (VHR) SAI. However, there lack such SGG datasets. Due to the complexity of large-size SAI, mining triplets < subject, relationship, object > heavily relies on long-range contextual reasoning. Consequently, SGG models designed for small-size natural imagery are not directly applicable to large-size SAI. This paper constructs a large-scale dataset for SGG in large-size VHR SAI with image sizes ranging from 512 × 768 to 27 860 × 31 096 pixels, named STAR (Scene graph generaTion in lArge-size satellite imageRy), encompassing over 210K objects and over 400K triplets. To realize SGG in large-size SAI, we propose a context-aware cascade cognition (CAC) framework to understand SAI regarding object detection (OBD), pair pruning and relationship prediction for SGG. We also release a SAI-oriented SGG toolkit with about 30 OBD and 10 SGG methods which need further adaptation by our devised modules on our challenging STAR dataset.
KW - Large-size satellite imagery
KW - object detection
KW - relationship prediction
KW - scene graph generation benchmark
UR - http://www.scopus.com/inward/record.url?scp=85202063078&partnerID=8YFLogxK
U2 - 10.1109/TPAMI.2024.3508072
DO - 10.1109/TPAMI.2024.3508072
M3 - 文章
AN - SCOPUS:85202063078
SN - 0162-8828
VL - 47
SP - 1832
EP - 1849
JO - IEEE Transactions on Pattern Analysis and Machine Intelligence
JF - IEEE Transactions on Pattern Analysis and Machine Intelligence
IS - 3
ER -