Journal Article

·2025

Scalable Remote Sensing Image Change Captioning using In-Context Learning

Berkay Ateş YTU , Orkhan Karimli YTU , Mehmet Fatih Amasyalı YTU , Ali Can Karaca YTU

Abstract

Remote Sensing Image Change Captioning (RSICC) plays a crucial role in detecting changes between bitemporal remote sensing images and generating descriptive captions that explain these alterations. Conventional approaches to developing such a successful system for RSICC tasks require substantial computational power, finely gathered images and captioning. In contrast, this study explores the advantages of In-Context Learning (ICL) combined with Visual Language Models for RSICC tasks to generate precise and accurate captions. The ICL approach involves presenting a few of demonstrations to the model with the queries to elicit expected outputs. Using the ICL, approximately 9× performance improvement is achieved on the LEVIR-CC dataset. Code available at https://github.com/ChangeCapsInRS/ICL-CC.

Keywords

Closed captioning Computer science Scalability Context (archaeology) Image (mathematics) Artificial intelligence Multimedia Computer vision Remote sensing Database Geology

Subject Areas

Multimodal Machine Learning Applications ·Computer Vision and Pattern Recognition ·Physical Sciences
Advanced Image and Video Retrieval Techniques ·Computer Vision and Pattern Recognition ·Physical Sciences
Video Surveillance and Tracking Methods ·Computer Vision and Pattern Recognition ·Physical Sciences