冯超然｜学术主页

🗂️ 代表性项目

* 共同一作 # 项目负责人

♠ 多模态生成 🤖 LingBot-Video: Scaling Mixture-of-Experts Video Pretraining for Embodied Intelligence 🎨 Style-GRPO: Semantic-Aware Preference Optimization for Image Style Transfer Guided by Reward Modeling 👹 Enhancing Spatial Understanding in Image Generation via Reward Modeling 🍟 WISE: A World Knowledge-Informed Semantic Evaluation for Text-to-Image Generation

🤖 LingBot-Video: Scaling Mixture-of-Experts Video Pretraining for Embodied Intelligence

Shuailei Ma, Jiaqi Liao, Xinyang Wang, Jingjing Wang, Chaoran Feng, Zijing Hu, Chong Bao, Zichen Xi, Yuqi Gan, Weisen Wang, Yanhong Zeng, Qin Zhao, Zifan Shi, Wei Wu, Hao Ouyang, Qiuyu Wang, Shangzhan Zhang, Jiahao Shao, Yipengjing Sun, Liangxiao Hu, Lunke Pan, Nan Xue, Kecheng Zheng, Yinghao Xu, Xing Zhu, Yujun Shen, Ka Leong Cheng

面向具身智能的 DiT-MoE 视频预训练范式；主要负责后训练奖励设计，以及针对物理合理性与任务完成度对齐的强化学习框架。

Technical Report 论文项目主页代码 ★629

🎨 Style-GRPO: Semantic-Aware Preference Optimization for Image Style Transfer Guided by Reward Modeling

Chaoran Feng^*,#, Jianbin Zhao*, Miao Yu* , Yingtao Li , Zhenyu Tang , Wangbo Yu , Yian Zhao , Xiaomin Li , Li Yuan† , Yonghong Tian†

一种由奖励建模引导、基于语义感知偏好优化的图像风格迁移框架。

CVPR 2026 论文

👹 Enhancing Spatial Understanding in Image Generation via Reward Modeling

Zhenyu Tang*, Chaoran Feng^*, Yufan Deng, Jie Wu, Xiaojie Li, Rui Wang, Yunpeng Chen, Daquan Zhou

一种通过奖励建模增强图像生成空间理解能力的框架。

CVPR 2026 论文项目主页代码 ★85

🍟 WISE: A World Knowledge-Informed Semantic Evaluation for Text-to-Image Generation

Yuwei Niu, Munan Ning, Mengren Zheng, Weiyang Jin, Bin Lin, Peng Jin, Jiaqi Liao, Chaoran Feng, Fanqing Meng, Kunpeng Ning, Bin Zhu, Li Yuan

首个面向文生图模型世界知识语义理解的评测基准，包含 1,000 条提示词，覆盖文化常识、时空推理与自然科学。

ICML 2026 论文代码 ★210

♠ 三维视觉 👾 EA3D: Event-Augmented 3D Diffusion for Generalizable Novel View Synthesis 🌫️ DeblurNVS: Geometric Latent Diffusion for Novel View Synthesis from Sparse Motion-Blurred Images 🍿 Breaking the Vicious Cycle: Coherent 3D Gaussian Splatting from Sparse and Motion-Blurred Views 🎨 Tune-Your-Style: Intensity-tunable 3D Style Transfer with Gaussian Splatting 🔥 Cycle3D: High-quality and Consistent Image-to-3D Generation via Generation-Reconstruction Cycle 🌀 NeuralGS: Bridging Neural Fields and 3D Gaussian Splatting for Compact 3D Representations 👤 NOFA++: Tuning-free NeRF-based One-shot Facial Avatar Reconstruction

👾 EA3D: Event-Augmented 3D Diffusion for Generalizable Novel View Synthesis

Wangbo Yu*, Chaoran Feng^*, Jianing Li, Aofan Zhang, Zhenyu Tang, Li Yuan† , Yonghong Tian†

一种事件增强的三维扩散框架，可根据事件流与稀疏 RGB 输入实现具有泛化能力的新视角合成。

ICLR 2026

🌫️ DeblurNVS: Geometric Latent Diffusion for Novel View Synthesis from Sparse Motion-Blurred Images

Changyue Shi, Wangbo Yu, Chaoran Feng, Li Yuan†

一种几何潜空间扩散框架，无需逐场景优化，即可从稀疏运动模糊图像直接合成高保真新视角。

arXiv 2026 论文代码 ★15

🍿 Breaking the Vicious Cycle: Coherent 3D Gaussian Splatting from Sparse and Motion-Blurred Views

Chaoran Feng, Zhankuo Xu, Yingtao Li, Jianbin Zhao, Jiashu Yang, Wangbo Yu, Li Yuan†, Yonghong Tian†

面向稀疏运动模糊视图的连贯三维高斯重建框架，将物理感知去模糊先验与扩散驱动的几何补全相结合。

IEEE T-CSVT 论文代码 ★98

🎨 Tune-Your-Style: Intensity-tunable 3D Style Transfer with Gaussian Splatting

Yian Zhao, Rushi Ye, Ruochong Zheng, Zesen Cheng, Chaoran Feng, Jiashu Yang, Pengchong Qiao, Chang Liu, Jie Chen†

一种基于三维高斯泼溅、支持强度调节的三维风格迁移框架。

ICCV 2025 论文

🔥 Cycle3D: High-quality and Consistent Image-to-3D Generation via Generation-Reconstruction Cycle

Zhenyu Tang*, Junwu Zhang*, Xinhua Cheng, Wangbo Yu, Chaoran Feng, Yatian Pang, Bin Lin, Li Yuan†

通过生成—重建循环统一扩散过程，实现高质量且一致的图生三维生成。

AAAI 2025 论文代码 ★219

🌀 NeuralGS: Bridging Neural Fields and 3D Gaussian Splatting for Compact 3D Representations

Zhenyu Tang*, Chaoran Feng*, Xinhua Cheng, Wangbo Yu, Junwu Zhang, Yuan Liu†, Xiaoxiao Long, Wenping Wang, Li Yuan†

使用神经场与轻量 MLP 紧凑编码三维高斯，适用于大规模场景表达。

AAAI 2026 论文代码 ★183

👤 NOFA++: Tuning-free NeRF-based One-shot Facial Avatar Reconstruction

Wangbo Yu, Chaoran Feng, Li Yuan†, and Yonghong Tian†

无需微调的单图 NeRF 人脸化身重建方法，可实现高保真重建与动态驱动。

IEEE T-CSVT 论文

♠ 神经形态视觉 🍡 GS2E: Gaussian Splatting is an Effective Data Generator for Event Stream Generation 🎈 E-4DGS: High-Fidelity Dynamic Reconstruction from the Multi-view Event Cameras ✨ EvaGaussians: Event Assisted Gaussian Splatting from Blurry Images ⚡ AE-NeRF: Augmenting Event-Based Neural Radiance Fields for Non-ideal Conditions and Larger Scenes

🍡 GS2E: Gaussian Splatting is an Effective Data Generator for Event Stream Generation

Yuchen Li*, Chaoran Feng^*,#, Zhenyu Tang, Kaiyuan Deng, Wangbo Yu, Yonghong Tian†, Li Yuan†

面向事件流生成的大规模数据集与新型仿真管线，以三维高斯泼溅作为高效数据生成器。

NeurIPS 2025 D&B Track 论文代码 ★20 数据集