FSS-1000

The FSS-1000 dataset is a benchmark used to evaluate visual in-context learning methods, specifically focusing on tasks involving foreground segmentation and object detection across diverse label spaces.

Papers using FSS-1000 (3)

REBASE: Reference-Background Subspace Elimination for Training-Free In-Context Segmentation2026MARS: a Multimodal Alignment and Ranking System for Few-Shot Segmentation2025CRCNet: Few-shot Segmentation with Cross-Reference and Region-Global Conditional Networks2022 · 27 cites