Seeking Consensus: Geometric-Semantic On-the-Fly Recalibration for Open-Vocabulary Remote Sensing Semantic Segmentation
For remote sensing image segmentation, this work addresses the static inference limitation of existing open-vocabulary models by providing a training-free recalibration method that improves performance across diverse scenes.
SeeCo is a plug-and-play framework that recalibrates open-vocabulary semantic segmentation models on-the-fly for remote sensing images, using geometric and semantic consensus learning to reduce semantic ambiguity and improve foreground activation. It achieves consistent gains across eight benchmarks without requiring training.
Open-vocabulary semantic segmentation (OVSS) in remote sensing images is a promising task that employs textual descriptions for identifying undefined land cover categories. Despite notable advances, existing methods typically employ a static inference paradigm, overlooking the distinct distribution of each scene, resulting in semantic ambiguity in diverse land covers and incomplete foreground activation. Motivated by this, we propose Seeking Consensus, termed SeeCo, a plug-and-play framework to boost the performance of training-free OVSS models in remote sensing images, which recalibrates arbitrary OVSS models on-the-fly by seeking dual consensus: geometric consensus learning (GCL) through multi-view consistent observations and semantic consensus learning (SCL) via textual description adaptive calibration, which assists collaborative recalibration of visual and textual semantics. The two consensus are injected via an online consensus injector (OCI), effectively alleviating the under-activation and semantic bias. SeeCo requires no specific training process, yet recalibrates semantic-geometric alignment for each unique scene during inference. Extensive experiments on eight remote sensing OVSS benchmarks show consistent gains, proving its effectiveness and universality.