Yiming Xie 谢一鸣


yimingxie@zju.edu.cn
ymxyimingxie@gmail.com
Research Assistant, ZJU
Hangzhou, China
GitHub

Latest News

  • [Apr 29, 2021] 1 paper is accepted to TPAMI'21
  • [Feb 29, 2021] 1 paper is accepted to CVPR'21
  • [Feb 24, 2020] 1 paper is accepted to CVPR'20

About Me


I'm currently working as Research Assistant in State Key Lab of CAD & CG, Zhejiang University supervised by Prof. Xiaowei Zhou. I received my Bachelor's degree in 2019 from Zhejiang University(ZJU).

Preprints



You Don’t Only Look Once: Constructing Spatial-Temporal Memory for Integrated 3D Object Detection and Tracking
Yiming Xie*, Jiaming Sun*, Siyu Zhang, Linghao Chen, Guofeng Zhang, Hujun Bao, Xiaowei Zhou. (*-indicates equal contributions)
In submission.

Publications



NeuralRecon: Real-Time Coherent 3D Reconstruction from Monocular Video
Yiming Xie*, Jiaming Sun*, Linghao Chen, Hujun Bao, Xiaowei Zhou. (* equal contributions)
Computer Vision and Pattern Recognition (CVPR), 2021.
Oral Presentation and Best Paper Candidate
[PDF][Code][Website]

Shape Prior Guided Instance Disparity Estimation for 3D Object Detection
Linghao Chen*, Jiaming Sun*, Yiming Xie, Siyu Zhang, Qing Shuai, Qinhong Jiang, Guofeng Zhang, Hujun Bao, Xiaowei Zhou. (* equal contributions)
Transactions on Pattern Analysis and Machine Intelligence (TPAMI), 2021.
[PDF][Code]

Stereo 3D Object Detection via Shape Prior GuidedInstance Disparity Estimation
Jiaming Sun*, Linghao Chen*, Yiming Xie, Siyu Zhang, Qinhong Jiang, Xiaowei Zhou, Hujun Bao. (* equal contributions)
Computer Vision and Pattern Recognition (CVPR), 2020.
[PDF][Code]

Talks & Presentations


  • [06/2020] PatchMatch in Multi-View Stereo. (PPTX   PDF   Video)
      @ ZJU-3DV Group
  • [04/2021] Real-Time 3D Reconstruction from Monocular Video. (In Chinese) (PDF   Video)
      @ AI Time

Projects



3D Scene Semantic Modeling

Deep Scene Flow Estimation with Iterative Soft-argmin and Edge-aware Regression
Proposed an edge-aware policy to improve the over-smoothing problem and utilized contrastive loss to learn a more discriminative descriptor.
Proposed an iterative soft argmin to cope with the multi-modal distributions of cost volume.