Shizhan Zhu


Email: zhshzhutah2 [at]

I am currently a PhD student in Berkeley AI Research Lab (BAIR), UC Berkeley, advised by Prof. Trevor Darrell.

Before that, I obtained my M.S. degree from MMLAB, CUHK. I was very fortunated to be advised by Prof. Chen Change Loy and Prof. Xiaoou Tang. I was also lucky to have the chance to work with Prof. Dahua Lin. I obtained my B.S. degree from EE, Tsinghua University.


Be your own Prada: Dressing People in New Outfits via Textual Descriptions and Generative Adversarial Networks

Shizhan Zhu , Sanja Fidler, Raquel Urtasun, Dahua Lin, Chen Change Loy
IEEE Conference on International Conference on Computer Vision (ICCV), 2017.
[ Paper ] [ Supplementary Material ] [ Full Codes on Github ]

What's New:
- Re-dress a given person based on text - keeping identity and pose unchanged.
- The left illustrative figure: The rendered fashion image conditioned on the same person (the same row) or the same text (the same column).
- Decoupled generation process.
- GAN for semantic segmentation generation.

Deep Cascaded Bi-Network for Face Hallucination

Shizhan Zhu , Sifei Liu, Chen Change Loy, Xiaoou Tang
European Conference on Computer Vision (ECCV), 2016.
[ Paper ] [ Codes on Github ]

What's New:
- Automatic facial texture synthesis with the bi-network. Charming results.
- Guided deep face editing.
Unconstrained Face Alignment via Cascaded Compositional Learning

Shizhan Zhu , Cheng Li, Chen Change Loy, Xiaoou Tang
IEEE Conference on Computer Vision and Pattern Recognition (CVPR), 2016.
[ Paper ] [ Full codes ] [ Revised full AFLW annotation ]

What's New:
- Fast speed for arbitrary-view-capable face alignment (e.g. profile-view).
- A new flavor of random forest splitting machanism.
Face Alignment by Coarse-to-Fine Shape Searching

Shizhan Zhu , Cheng Li, Chen Change Loy, Xiaoou Tang
IEEE Conference on Computer Vision and Pattern Recognition (CVPR), 2015.
[ Paper ] [ Extended Abstract ] [ Demo Video ] [ Full Codes on Github ]
[ A Comprehensive Evaluation of Our Approach by IBUG ]

What's New:
- Address the initialization sensitivity problem of cascaded regression.
- State-of-the-art results on the 300W benchmark. Widely acknowledged implementation.

Seminar Talk

  • Traditional Face Alignment - Intuition and Applications [ Slides ]

Selected Honors and Awards

  • Distinguished graduate award of Tsinghua University (Cum Laude, top 10% in EE dept.), 2014.
  • Distinguished bachelor thesis award of Tsinghua University (12 out of 245 in EE dept.), 2014.
  • First class of scholarship for two years in Tsinghua, 2012, 2013.

Professional Activities

  • Conference Reviewer
    - CVPR (Computer Vision and Pattern Recognition);
    - ICCV (International Conference on Computer Vision);
  • Journal Reviewer
    - TPAMI (Trasactions on Pattern Analysis and Machine Intelligence)
    - CVIU (Computer Vision and Image Understanding);
    - IET Computer Vision.

Published with GitHub Pages