STAND: Semantic Anchoring Constraint with Dual-Granularity Disambiguation for Remote Sensing Image Change Captioning

TL;DR

STAND introduces a novel framework for remote sensing image change captioning that progressively resolves viewpoint, scale, and knowledge ambiguities using semantic anchoring and dual-granularity disambiguation.

Contribution

It proposes a comprehensive method combining interpretability, dual-granularity disambiguation, and semantic anchoring to improve change captioning accuracy in remote sensing images.

Findings

STAND outperforms existing methods in accuracy on benchmark datasets.

The dual-granularity disambiguation effectively resolves spatial uncertainties.

Semantic concept anchoring improves the translation of features into precise captions.

Abstract

Remote sensing image change captioning (RSICC) aims to describe the difference between two remote sensing images. While recent methods have explored video modeling, they largely overlook the inherent ambiguities in viewpoint, scale, and prior knowledge, lacking effective constraints on the encoder. In this paper, we present STAND, a Semantic Anchoring Constraint with Dual-Granularity Disambiguation for RSICC, to progressively resolve these ambiguities. Specifically, to establish a reliable feature foundation, we first introduce an interpretable constraint to regularize temporal representations. Operating on these purified features, a dual-granularity disambiguation module resolves spatial uncertainties by coupling macro-level global context aggregation for viewpoint confusion with micro-level frequency-refocused attention for small-object scale enhancement. Ultimately, to translate these visually disambiguated features into precise text, a semantic concept anchoring module leverages language categorical priors to tackle knowledge ambiguity during decoding. Extensive experiments verify the superiority of STAND and its effectiveness in addressing ambiguities.