PDAR-RSITR: A Progressive Decoupling-Aggregation-Refinement Framework for Remote Sensing Image-Text Retrieval

Wang, Shuhuai; Pei, Songwei; Liu, Bingfeng; Chai, Duo; Huang, Yuanzhou; Liu, Jia; Li, Qian; Wang, Shangguang

doi:10.24963/ijcai.2026/661

PDAR-RSITR: A Progressive Decoupling-Aggregation-Refinement Framework for Remote Sensing Image-Text Retrieval

Shuhuai Wang, Songwei Pei, Bingfeng Liu, Duo Chai, Yuanzhou Huang, Jia Liu, Qian Li, Shangguang Wang

Download BibTeX Cite

Proceedings of the Thirty-Fifth International Joint Conference on Artificial Intelligence. Main Track. Pages 5937-5945.

https://doi.org/10.24963/ijcai.2026/661

Keywords

Natural Language Processing: Information retrieval and text mining

Abstract

Remote Sensing Image-Text Retrieval (RSITR) aims to achieve precise retrieval between remote sensing images and textual descriptions. However, existing methods neglect the multi-dimensional cognitive attributes inherent in remote sensing data and struggle to handle them simultaneously, leading to suboptimal retrieval performance. In this paper, we propose a novel Progressive Decoupling-Aggregation-Refinement framework for RSITR (PDAR-RSITR) to comprehensively capture multi-dimensional cognitive attributes. Specifically, we adapt Multi-dimensional Cognitive Decoupling to learn features across cognitive attributes of different dimensions via multi-process clustering. Subsequently, we utilize Salient Cognitive Aggregation to select and aggregate the most salient attributes via dynamic routing. Furthermore, we propose Expert Collaborative Refinement to enhance critical cross-modal relationships via three complementary expert perspectives. Extensive experimental results demonstrate that PDAR-RSITR significantly outperforms existing state-of-the-art methods across multiple metrics.

Shuhuai Wang, Songwei Pei, Bingfeng Liu, Duo Chai, Yuanzhou Huang, Jia Liu, Qian Li, and Shangguang Wang. 2026. PDAR-RSITR: A Progressive Decoupling-Aggregation-Refinement Framework for Remote Sensing Image-Text Retrieval. In Proceedings of the Thirty-Fifth International Joint Conference on Artificial Intelligence, IJCAI-26. IJCAI, 5937–5945. https://doi.org/10.24963/ijcai.2026/661

@inproceedings{ijcai2026p661,
  title     = {PDAR-RSITR: A Progressive Decoupling-Aggregation-Refinement Framework for Remote Sensing Image-Text Retrieval},
  author    = {Wang, Shuhuai and Pei, Songwei and Liu, Bingfeng and Chai, Duo and Huang, Yuanzhou and Liu, Jia and Li, Qian and Wang, Shangguang},
  booktitle = {Proceedings of the Thirty-Fifth International Joint Conference on
               Artificial Intelligence, {IJCAI-26}},
  publisher = {International Joint Conferences on Artificial Intelligence Organization},
  editor    = {Diego Calvanese},
  pages     = {5937--5945},
  year      = {2026},
  month     = {8},
  note      = {Main Track},
  doi       = {10.24963/ijcai.2026/661},
  url       = {https://doi.org/10.24963/ijcai.2026/661},
}

TY  - CPAPER
TI  - PDAR-RSITR: A Progressive Decoupling-Aggregation-Refinement Framework for Remote Sensing Image-Text Retrieval
AU  - Wang, Shuhuai
AU  - Pei, Songwei
AU  - Liu, Bingfeng
AU  - Chai, Duo
AU  - Huang, Yuanzhou
AU  - Liu, Jia
AU  - Li, Qian
AU  - Wang, Shangguang
T2  - Proceedings of the Thirty-Fifth International Joint Conference on Artificial Intelligence, IJCAI-26
PB  - International Joint Conferences on Artificial Intelligence Organization
PY  - 2026
SP  - 5937
EP  - 5945
DO  - 10.24963/ijcai.2026/661
UR  - https://doi.org/10.24963/ijcai.2026/661
ER  -