ICSI Vision Group
UC Berkeley
1947 Center Street
Berkeley, CA 94704
E-mail: qianyu at icsi.berkeley.edu
[Github] [Google Scholar]

I'm a postdoctoral research fellow at ICSI/UC Berkeley with Prof. Stella Yu. My research focuses on sketch understanding including sketch recognition and sketch-based image retrieval. I am also interested in human visual perception.

I received my Ph.D. degree (2014-2018) from QMUL/SketchX Lab, advised by Dr. Yi-Zhe Song and Prof. Tao Xiang. I graduated from BUPT with a B.S. degree in Telecommunications Engineering with Management.


  1. New!SketchyScene: Richly-Annotated Scene Sketches
    Changqing Zou*, Qian Yu*, Ruofei Du, Haoran Mo, Yi-Zhe Song, Tao Xiang, Chengying Gao, Baoquan Chen and Hao Zhang
    European Conference on Computer Vision (ECCV), 2018 
    [PDF]   [Supplementary]   [Dataset]  


  2. Fine-grained Instance-level Sketch-based Image Retrieval
    Qian Yu, Jifei Song, Yi-Zhe Song, Tao Xiang and Timothy M. Hospedales
    2018 


  3. Deep Spatial-Semantic Attention for Fine-Grained Sketch-Based Image Retrieval
    Jifei Song*, Qian Yu*, Yi-Zhe Song, Tao Xiang and Timothy M. Hospedales
    International Conference on Computer Vision (ICCV), 2017 
    [PDF]   [Supplementary]   [Model(Tensorflow, 106.1M)]   [Bibtex]  [Project Page]  


  4. The Devil is in the Middle: Exploiting Mid-level Representations for Cross-Domain Instance Matching
    Qian Yu, Xiaobin Chang, Yi-Zhe Song, Tao Xiang and Timothy M. Hospedales
    arXiv:1711.08106, 2017 
    [PDF]   [Code [Pytorch, ResNet50M][TensorFlow, coming soon]]   [Bibtex


  5. Sketch-a-Net: a Deep Neural Network that Beats Humans
    Qian Yu, Yongxin Yang, Feng Liu, Yi-Zhe Song, Tao Xiang and Timothy M. Hospedales
    International Journal of Computer Vision (IJCV), May 2017
    [PDF]   [DOI]   [Model (Caffe&Tensorflow, 68.1M)]   [Bibtex


  6. Sketch Me that Shoe
    Qian Yu, Feng Liu, Yi-Zhe Song, Tao Xiang, Timothy M. Hospedales, Chen Change Loy
    IEEE Conference on Computer Vision and Pattern Recognition (CVPR), 2016 (Oral)
    [PDF]   [Slides]   [Model (Caffe, 297M)]   [Model (Tensorflow, 99M)]   [Demo]   [Bibtex]  [Project Page]  


  7. Sketch-a-Net that Beats Humans
    Qian Yu*, Yongxin Yang*, Yi-Zhe Song, Tao Xiang, Timothy M. Hospedales
    British Machine Vision Conference (BMVC), 2015 (Best Science Paper)
    [PDF]   [Slides]   [Code]   [Bibtex]  


QMUL FG-SBIR database

  • Shoe/Chair-v1
  • Shoe-v2
  • Handbags
    1. QMUL Postgraduate Research Fund 2017
    2. ICCV Student Volunteer 2017
    3. BMVA Travel Bursary 2015
    4. QMUL-CSC PhD Scholarship 2014-2018
    5. Outstanding Graduate of BUPT 2014
  • Sketch Me that Shoe at CVPR'16  
  • Sketch-a-Net that Beats Humans at BMVC'15  
  • Teaching assistant at QMUL for the following courses:
  • ECS416U - Introduction to Multimedia
  • ECS705P - Software and Network Services Design
  • Conference reviewer for CVPR'18, SIGGRAPH'18, BMVC'18
    Journal reviewer for International Journal of Computer Vision (IJCV), Transactions on Image Processing (TIP), Neurocomputing, Computer Vision and Image Understanding (CVIU), Computer&Graphics (CAG), The Visual Computer
    I am a STEM Ambassador. In my spare time, I often volunteer in children's science education activities. I enjoy playing with children and hope to inspire their interest in STEM subjects!