日月光集团 · 相关文章
相关文章
Objective-Aligned Direct Answer SFT for Robust Multi-Frame Medical VQA
ArXiv CS.CV2026-07-31
MPIE-Bench: Benchmarking Anatomically Plausible Multi-Person Interaction Editing
ArXiv CS.CV2026-07-31
BlindPSNR: A No-Reference Fidelity Predictor for Low-Light Image Enhancement
ArXiv CS.CV2026-07-31
4DHumanDiff: Direct Text-to-4DGS Generation for Consistent 360-Degree Dynamic Humans
ArXiv CS.CV2026-07-31
RefineSVG: Visual Feedback-Driven Reinforcement Learning for Image-to-SVG Generation
ArXiv CS.CV2026-07-31