Education
Northwestern Polytechnology Universtiy
Ph. D in School of Automation, advised by Prof. Dingwen Zhang
Ph. D in School of Automation, advised by Prof. Dingwen Zhang
Fall 2022 - So far (don't ask)
Beijing University of Chemical Technology
B.S. in School of Information and Technology.
B.S. in School of Information and Technology.
Fall 2018 - Spring 2022
Employment
Baidu, Inc.
Research Intern
Research Intern
Dec 2023 - Now
Beijing, CN
Beijing, CN
Zhejiang Lab
Research Intern
Research Intern
Jun 2022 - Jun 2023
Hangzhou, CN
Hangzhou, CN
Netease, Inc.
Research Intern
Research Intern
Sep 2021 - Jun 2022
Beijing, CN
Beijing, CN
Institute of Automation, Chniese Academy of Sciences
Summer Research Intern
Summer Research Intern
Jul 2019 - Sep 2019
Beijing, CN
Beijing, CN
Publications
XLD: A Cross-Lane Dataset for Benchmarking Novel Driving View Synthesis
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.
Arxiv, 2024
VDG: Vision-Only Dynamic Gaussian for Driving Simulation
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.
Arxiv, 2024
GGRt: Towards Pose-free Generalizable 3D Gaussian Splatting in Real-time
As the first pose-free generalizable 3D-GS framework, GGRt achieves inference at > 5 FPS and real-time rendering at > 100 FPS
ECCV, 2024
GP-NeRF: Generalized Perception NeRF for Context-Aware 3D Scene Understanding
GP-NeRF achieves remarkable performance improvements for instance and semantic segmentation in both synthesis and real-world datasets.
CVPR, 2024
LTGC: Long-Tail Recognition via Leveraging LLMs-driven Generated Content
We propose a novel generative and fine-tuning framework, LTGC, to handle long-tail recognition via leveraging generated content.
CVPR, 2024
Weakly Supervised Semantic Segmentation via Alternate Self-Dual Teaching
We build a novel end-to-end learning framework, alternate self-dual teaching (ASDT), based on a dual-teacher single-student network architecture.
IEEE Transaction of Image Processing, 2024
Boosting low-data instance segmentation by unsupervised pre-training with saliency prompt
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.
CVPR, 2023