FarSLIP: Discovering Effective CLIP Adaptation for Fine-Grained Remote Sensing Understanding

TL;DR

FarSLIP introduces a novel fine-grained RS image-text dataset and a new CLIP adaptation method that enhances spatial awareness and semantic coherence, significantly improving performance on RS vision-language tasks.

Contribution

It constructs the first multi-granularity RS dataset with object-level supervision and proposes FarSLIP, a new fine-grained CLIP tuning framework utilizing patch-to-patch distillation and CLS token-based region-category alignment.

Findings

Sets new state-of-the-art on RS semantic segmentation

Improves zero-shot classification accuracy in RS domain

Enhances image-text retrieval performance

Abstract

As CLIP's global alignment limits its ability to capture fine-grained details, recent efforts have focused on enhancing its region-text alignment. However, current remote sensing (RS)-specific CLIP variants still inherit this limited spatial awareness. We identify two key limitations behind this: (1) current RS image-text datasets generate global captions from object-level labels, leaving the original object-level supervision underutilized; (2) despite the success of region-text alignment methods in general domain, their direct application to RS data often leads to performance degradation. To address these, we construct the first multi-granularity RS image-text dataset, MGRS-200k, featuring rich object-level textual supervision for RS region-category alignment. We further investigate existing fine-grained CLIP tuning strategies and find that current explicit region-text alignment methods, whether in a direct or indirect way, underperform due to severe degradation of CLIP's semantic coherence. Building on these, we propose FarSLIP, a Fine-grained Aligned RS Language-Image Pretraining framework. Rather than the commonly used patch-to-CLS self-distillation, FarSLIP employs patch-to-patch distillation to align local and global visual cues, which improves feature discriminability while preserving semantic coherence. Additionally, to effectively utilize region-text supervision, it employs simple CLS token-based region-category alignment rather than explicit patch-level alignment, further enhancing spatial awareness. FarSLIP features improved fine-grained vision-language alignment in RS domain and sets a new state of the art not only on RS open-vocabulary semantic segmentation, but also on image-level tasks such as zero-shot classification and image-text retrieval. Our dataset, code, and models are available at https://github.com/NJU-LHRS/FarSLIP.