Unifying Heterogeneous Multi-Modal Remote Sensing Detection Via Language-Pivoted Pretraining
Lists
Only you see this
Abstract
Heterogeneous multi-modal remote sensing object detection aims to accurately detect objects from diverse sensors (e.g., RGB, SAR, Infrared). Existing approaches largely adopt a late alignment paradigm, in which modality alignment and task-specific optimization are entangled during downstream fine-tuning. This tight coupling complicates optimization and often results in unstable training and suboptimal generalization. To address these limitations, we propose BabelRS, a unified language-pivoted pretraining framework that explicitly decouples modality alignment from downstream task learning. BabelRS comprises two key components: Concept-Shared Instruction Aligning (CSIA) and Layerwise Visual-Semantic Annealing (LVSA). CSIA aligns each sensor modality to a shared set of linguistic concepts, using language as a semantic pivot to bridge heterogeneous visual representations. To further mitigate the granularity mismatch between high-level language representations and dense detection objectives, LVSA progressively aggregates multi-scale visual features to provide fine-grained semantic guidance. Extensive experiments demonstrate that BabelRS stabilizes training and consistently outperforms state-of-the-art methods without bells and whistles. Code: \href{https://github.com/zcablii/SM3Det}{github.com/zcablii/SM3Det}.
BibTeX
@inproceedings{li2026unifying,
author = {Yuxuan Li and Yuming Chen and Yunheng Li and Ming-Ming Cheng and Xiang Li and jian Yang},
title = {{Unifying Heterogeneous Multi-Modal Remote Sensing Detection Via Language-Pivoted Pretraining}},
booktitle = {International Conference on Machine Learning},
year = {2026},
url = {https://openreview.net/forum?id=Afb6Yo2rIf},
}Loading related content...