CPPMN

01 — Overview

Overview

Remote-sensing scenes can be described from multiple geographic perspectives. Retrieval systems that collapse those perspectives into one representation may match a query to the wrong region or overlook the relevant spatial relationship.

CPPMN progressively learns full-image perspectives, exposes perspective-specific cross-modal relationships, and then aligns image and language features semantically.

02 — Contributions

Key Contributions

• 01

  Uses positive text descriptions to supervise full-perspective visual feature learning.

• 02

  Transforms implicit perspective features into explicit cross-modal relationship graphs.

• 03

  Applies cascaded Transformer layers for progressive image–text semantic alignment.

03 — Method

Method

The network combines a compensation module for full-perspective modeling, a graph transformation module for locating individual perspectives, and a cascaded Transformer for cross-modal semantic alignment. Graph density and connectivity help identify the perspective referred to by the query.

04 — Evaluation

Results

Quantitative and qualitative experiments across four remote-sensing image–text retrieval datasets demonstrate the value of progressive perspective matching and semantic alignment.

05 — Reference

Citation

@Article{Zheng_2025_CPPMN,
  author  = {Zheng, Chengyu and Li, Xiu and Liang, Xinyue and Huang, Lei and Du, Shan and Nie, Jie and Dong, Junyu},
  title   = {Cross-Modal Progressive Perspective Matching Network for Remote Sensing Image-Text Retrieval},
  journal = {IEEE Transactions on Multimedia},
  year    = {2025},
  volume  = {27},
  pages   = {3966--3978},
  doi     = {10.1109/TMM.2025.3535365}
}