Hello, I'm Hao Li (Leo Li)

I’m a Second-year (2022-) Ph.D. student in the BRAIN Lab, Northwestern Polytechnology University (NWPU), supervised by Prof. Dingwen Zhang and Prof. Junwei Han (IEEE Fellow). Currently, I am a Reaserch Intern in the Department of Computer Vision Technology (VIS), Baidu Inc., advised by Dr. Chenming Wu and Dr. Jingdong Wang (IEEE Fellow). My research area lies in the field of Semi-Supervised Learning, 3D Vision and Semantic Understanding.

CV Scholar Github

News

Jul 2024: GGRt has been accepted to ECCV 2024 (CCF B).
Jun 2024: Invited talk on 3D视觉工坊.
Feb 2024: GP-NeRF (Hightlight) and LTGC (Oral) have been accepted to CVPR 2024 (CCF A).
Dec 2023: Joining the Department of Computer Vision Technology (VIS), Baidu Inc. as Research Intern.
Oct 2023: ASDT has been accepted to TIP 2023 (SCI Q1).
Feb 2023: Saliency Prompt has been accepted to CVPR 2023 (CCF A).
Jun 2022: Joining Zhejiang Lab as Research Intern.

Publications

XLD: A Cross-Lane Dataset for Benchmarking Novel Driving View Synthesis

Hao Li, Chenming Wu, , , Errui Ding, Dingwen Zhang, Jingdong Wang

Arxiv, 2024

This paper presents a novel driving view synthesis dataset and benchmark specifically designed for autonomous driving simulations. This dataset is unique as it includes testing images captured by deviating from the training trajectory by 1-4 meters.

Project Page PDF arXiv

VDG: Vision-Only Dynamic Gaussian for Driving Simulation

Hao Li, Jingfeng Li, Dingwen Zhang, Chenming Wu, , , , Errui Ding, Jingdong Wang, Junwei Han

Arxiv, 2024

This paper addresses this issue by integrating self-supervised VO into our pose-free dynamic Gaussian method (VDG) to boost pose and depth initialization and static-dynamic decomposition.

Project Page PDF arXiv

GGRt: Towards Pose-free Generalizable 3D Gaussian Splatting in Real-time

Hao Li, Yuanyuan Gao, Chenming Wu, Dingwen Zhang, Yalun Dai, , , Errui Ding, Jingdong Wang, Junwei Han

ECCV, 2024

As the first pose-free generalizable 3D-GS framework, GGRt achieves inference at > 5 FPS and real-time rendering at > 100 FPS

Project Page PDF arXiv

GP-NeRF: Generalized Perception NeRF for Context-Aware 3D Scene Understanding

Hao Li, Dingwen Zhang, Yalun Dai, Nian Liu, Lechao Cheng, Jingfeng Li, Jingdong Wang, Junwei Han

CVPR, 2024 Highlight

GP-NeRF achieves remarkable performance improvements for instance and semantic segmentation in both synthesis and real-world datasets.

Project Page PDF arXiv Code

LTGC: Long-Tail Recognition via Leveraging LLMs-driven Generated Content

Qihao Zhao, Yalun Dai, Hao Li, Wei Hu, Fan Zhang, Jun Liu

CVPR, 2024 Oral Presentation

We propose a novel generative and fine-tuning framework, LTGC, to handle long-tail recognition via leveraging generated content.

Project Page PDF arXiv Code

Weakly Supervised Semantic Segmentation via Alternate Self-Dual Teaching

Dingwen Zhang, Hao Li, Chaowei Fang, Lechao Cheng, Mingming Cheng, Junwei Han

IEEE Transaction of Image Processing, 2024

We build a novel end-to-end learning framework, alternate self-dual teaching (ASDT), based on a dual-teacher single-student network architecture.

Project Page PDF arXiv Code

Boosting low-data instance segmentation by unsupervised pre-training with saliency prompt

Hao Li, Dingwen Zhang, Nian Liu, Lechao Cheng, Yalun Dai, Xinggang Wang, Junwei Han

CVPR, 2023

Inspired by the recent success of the Prompting technique, we introduce a new pre-training method that boosts QEIS models by giving Saliency Prompt for queries/kernels.

Project Page PDF arXiv