Test-Time Adaptation for Height Completion via Self-Supervised ViT Features and Monocular Foundation Models

Test-Time Adaptation for Height Completion via Self-Supervised ViT Features and Monocular Foundation Models explores A test-time adaptation framework for accurate digital surface model completion using foundation models, reducing reconstruction error by up to 46% without task-specific training.. Commercial viability score: 7/10 in Geospatial AI.