About Me

I am a third-year master's student in Control Science and Engineering at Northeastern University, advised by Prof. Qi Wang (Changjiang Scholar).

My research focuses on 3D vision and embodied intelligence, especially scalable 3D perception, scene completion, and robot manipulation.

Recent focus: I am working on World Action Model research for embodied algorithms, with an emphasis on robot manipulation strategies based on tactile-physical perception fusion.

News

2026/02
PatchScene was accepted to CVPR 2026.
2026/01
VAP-SAM was accepted by IEEE Transactions on Instrumentation and Measurement.
2025/05
EM-DETR was accepted by IEEE Sensors Journal.

Publications (* Equal contribution)

PatchScene: Patch-based Voxel Diffusion for Large-Scale Scene Completion

CVPR 2026 Co-first Author
Qingdong Xu*, Jiajun Zhu*, Shilin Zhu*, Xinjing He, Chao Lu, Huanran Wang, Jiyao Zhang
Paper Code

A diffusion framework for large-scale LiDAR scene completion. PatchScene generates fine-grained geometry in local voxel patches, fuses spatial and temporal context, and expands completion through an annular-flow strategy for scalable outdoor reconstruction. 

@inproceedings{xu2026patchscene,
  title={PatchScene: Patch-based Voxel Diffusion for Large-Scale Scene Completion},
  author={Xu, Qingdong and Zhu, Jiajun and Zhu, Shilin and He, Xinjing and Lu, Chao and Wang, Huanran and Zhang, Jiyao},
  booktitle={Conference on Computer Vision and Pattern Recognition},
  year={2026}
}

VAP-SAM: Prompt-Free SAM for Insubstantial Gas Segmentation

IEEE TIM 2026 JCR Q1
Shilin Zhu, Qi Wang, Yixuan Jing, Mengge Yang
Paper Demo

Builds an open-source gas segmentation dataset and replaces manual prompt engineering with a visual autonomous perception module. The method distills prompt-trained SAM priors into a prompt-free segmentation pipeline. 

@article{zhu2026vapsam,
  title={VAP-SAM: Prompt-Free SAM for Insubstantial Gas Segmentation},
  author={Zhu, Shilin and Wang, Qi and Jing, Yixuan and Yang, Mengge},
  journal={IEEE Transactions on Instrumentation and Measurement},
  volume={75},
  pages={1--9},
  year={2026},
  doi={10.1109/TIM.2026.3660401}
}

EM-DETR: Efficient Multiscale DETR for Gas Leak Detection Based on Infrared Imaging

IEEE Sensors Journal 2025 JCR Q1
Qi Wang, Shilin Zhu, Zhenlin Lu, Yunlong Sun, Yizhuo Qian
Paper

An efficient infrared gas leak detector that reduces redundant feature extraction, strengthens sensitivity to gas target spatial cues, and improves multi-scale feature fusion for small and diffuse gas regions. 

@article{zhu2025emdetr,
  title={EM-DETR: Efficient Multiscale DETR for Gas Leak Detection Based on Infrared Imaging},
  author={Wang, Qi and Zhu, Shilin and Lu, Zhenlin and Sun, Yunlong and Qian, Yizhuo},
  journal={IEEE Sensors Journal},
  volume={25},
  number={11},
  pages={20538--20546},
  year={2025},
  doi={10.1109/JSEN.2025.3558317}
}

Experience

Data Simulation and Embodied Intelligence WAM Algorithms

PaXini Tech, advised by Prof. Zhengxue Cheng
  • Retargeted hand-collected grasping data and used Isaac Lab to replay trajectories for LeRobot-style Pi0.5 training and real-world validation.
  • Pretrained a tactile encoder with MAE-style random masking to recover 3D point positions and tactile force signals.
  • Integrated tactile representations into Lingbot-VA with image and action denoising while preserving causal masking for future-frame isolation.
PaXini experience photo 1 PaXini experience photo 2
2025.12 - 2026.04

Honors & Awards

  • BYD Scholarship, 2026
  • First-class Scholarship, Northeastern University, 2023, 2024, 2025
  • Outstanding Graduate of Anhui Province, 2023
  • National First Prize, National University Student Intelligent Car Race, 2022
  • First Prize, Anhui Electronic Design Competition, 2022
  • National Scholarship, 2020