Style-GRPO
Semantic-Aware Preference Optimization for Image Style Transfer Guided by Reward Modeling
人工智能本科生 · 大连理工大学AI Undergraduate · Dalian University of Technology
你好,我是李英韬。我关注三维视觉、3D Gaussian Splatting 与图像风格迁移,并持续记录研究过程中的想法与实践。
Hi, I am Yingtao Li. I work on 3D vision, 3D Gaussian Splatting and image style transfer, while documenting ideas and experiments along the way.
我目前就读于大连理工大学人工智能专业,GPA 3.9/5.0,专业排名 11/312。研究兴趣集中在计算机视觉与三维场景表示。
I study Artificial Intelligence at Dalian University of Technology, with a GPA of 3.9/5.0 and a rank of 11/312. My interests center on computer vision and 3D scene representation.
目前有两篇论文分别被 CVPR 2026 与 IEEE T-CSVT 接收。也在探索 RGB 与事件相机融合的三维重建。
I have two papers accepted by CVPR 2026 and IEEE T-CSVT. I am also exploring RGB and event-camera fusion for 3D reconstruction.
Semantic-Aware Preference Optimization for Image Style Transfer Guided by Reward Modeling
Breaking the Vicious Cycle: Coherent 3D Gaussian Splatting from Sparse and Motion-Blurred Views
融合 RGB 与事件相机数据,探索更鲁棒的流式三维重建。Fusing RGB and event-camera data for more robust streaming 3D reconstruction.
这里是独立的个人 Blog,不与既有学习笔记混用。文章将由我陆续更新。This is an independent personal blog, separate from my existing study notes. Posts will be published here over time.
关于课程、科研、时间管理,以及如何在不完美的条件下继续向前。On coursework, research, time management, and moving forward under imperfect conditions.
从 2023 年的遗憾到 2024 年的重新出发,写给那个曾经从低处爬起来的自己。From the regrets of 2023 to starting over in 2024, a note to the self who once climbed back up.
人工智能专业,工学学士在读B.Eng. candidate in Artificial Intelligence
GPA 3.9/5.0 · Rank 11/312
两篇论文被接收Two papers accepted