CVPR 2026
Research Snapshot
Visual Evaluation & Generative Intelligence
My research lies at the intersection of Low-level Vision, AIGC, and MLLMs. I build frameworks like PosterCraft and PosterOmni for high-quality poster generation and editing, design reward models for human-aligned generation in PosterReward, and explore MLLM applications in image restoration tasks like SnowMaster and GenHaze.
4
First/Co-first Author
(CVPR*3, ICLR*1)
(CVPR*3, ICLR*1)
5
Total Publications
Highlight Research
CVPR 2026
CVPR 2026
PosterOmni — Generalized Artistic Poster Creation via Task Distillation and Unified Reward Feedback
One model for poster creation—unifying local edits and global design for generalized multi-task image/poster-to-poster generation. ✨ Your intelligent assistant for high-quality aesthetic poster creation!
ICLR 2026
Selected Publications
[CVPR 2026] PosterOmni — Generalized Artistic Poster Creation via Task Distillation and Unified Reward Feedback
One model for poster creation—unifying local edits and global design for generalized multi-task image/poster-to-poster generation.
[CVPR 2026] PosterReward: Unlocking Accurate Evaluation for High-Quality Graphic Design Generation
A comprehensive reward model for design aesthetics and typography, trained on automated preference dataset.
[ICLR 2026] PosterCraft — Rethinking High-Quality Aesthetic Poster Generation in a Unified Framework
A unified framework for high-quality aesthetic poster generation.
[ICCV 2025] GenHaze — Pioneering Controllable One-Step Realistic Haze Generation for Real-World Dehazing
Proposes a one-step, reference-controllable haze generator that better matches real-world haze complexity.
[CVPR 2025] SnowMaster — Comprehensive Real-world Image Desnowing via MLLM with Multi-Model Feedback Optimization
Uses an MLLM as an evaluator to rank/filter pseudo-labels, enabling semi-supervised training for real-world desnowing.