Back to Search
Start Over
A Lightweight Sparse Focus Transformer for Remote Sensing Image Change Captioning
- Source :
- IEEE Journal of Selected Topics in Applied Earth Observations and Remote Sensing 2024
- Publication Year :
- 2024
-
Abstract
- Remote sensing image change captioning (RSICC) aims to automatically generate sentences that describe content differences in remote sensing bitemporal images. Recently, attention-based transformers have become a prevalent idea for capturing the features of global change. However, existing transformer-based RSICC methods face challenges, e.g., high parameters and high computational complexity caused by the self-attention operation in the transformer encoder component. To alleviate these issues, this paper proposes a Sparse Focus Transformer (SFT) for the RSICC task. Specifically, the SFT network consists of three main components, i.e. a high-level features extractor based on a convolutional neural network (CNN), a sparse focus attention mechanism-based transformer encoder network designed to locate and capture changing regions in dual-temporal images, and a description decoder that embeds images and words to generate sentences for captioning differences. The proposed SFT network can reduce the parameter number and computational complexity by incorporating a sparse attention mechanism within the transformer encoder network. Experimental results on various datasets demonstrate that even with a reduction of over 90\% in parameters and computational complexity for the transformer encoder, our proposed network can still obtain competitive performance compared to other state-of-the-art RSICC methods. The code is available at \href{https://github.com/sundongwei/SFT_chag2cap}{Lite\_Chag2cap}.
- Subjects :
- Computer Science - Computer Vision and Pattern Recognition
Subjects
Details
- Database :
- arXiv
- Journal :
- IEEE Journal of Selected Topics in Applied Earth Observations and Remote Sensing 2024
- Publication Type :
- Report
- Accession number :
- edsarx.2405.06598
- Document Type :
- Working Paper
- Full Text :
- https://doi.org/10.1109/JSTARS.2024.3471625