Presents Reason50K dataset and ReasonBrain framework for hypothetical instruction-based image editing that requires physical, temporal, causal, and story reasoning.
Fireedit: Fine-grained instruction-based image editing via region-aware vision language model.arXiv preprint arXiv:2503.19839
1 Pith paper cite this work. Polarity classification is still indexing.
1
Pith paper citing it
fields
cs.CV 1years
2025 1verdicts
UNVERDICTED 1representative citing papers
citing papers explorer
-
Reasoning to Edit: Hypothetical Instruction-Based Image Editing with Visual Reasoning
Presents Reason50K dataset and ReasonBrain framework for hypothetical instruction-based image editing that requires physical, temporal, causal, and story reasoning.