A Benchmark for Robust GUI Grounding in Complex Multi-Window Desktop Environments
WinDeskGround is a reproducible benchmark toolkit for evaluating GUI grounding models in realistic multi-window desktop environments.
It includes:
- sensitivity-controlled experiments for occlusion, semantic distraction, and clutter
- simulation difficulty levels from L1 to L5
- unified evaluation scripts for multiple local or API-based models