Representing bounding-box coordinates as axis-specific hundreds, tens, and ones tokens, plus a geometry-aware GRPO reward, improves generative visual grounding accuracy.
Proceedings of the IEEE/CVF Winter Conference on Applications of Computer Vision , year=
1 Pith paper cite this work. Polarity classification is still indexing.
1
Pith paper citing it
fields
cs.CV 1years
2026 1verdicts
CONDITIONAL 1representative citing papers
citing papers explorer
-
Hi-Token: Hierarchical Coordinate Tokenization for Generative Visual Grounding
Representing bounding-box coordinates as axis-specific hundreds, tens, and ones tokens, plus a geometry-aware GRPO reward, improves generative visual grounding accuracy.