Welcome to My Homepage

Hello! I am a postdoc reseacher in University of Technology Sydney, working with Dr. Yi Yang. I was a postdoc in University of Texas at San Antonio working with Prof. Qi Tian. I was an intern in Microsoft Research mentored by Dr. Jingdong Wang. I received my PhD degree in 2015 from Department of Electronic Engineering, Tsinghua University, supervised by Prof. Shengjin Wang. I received my B.E from School of Life Sciences, Tsinghua University in 2010. I am interested in deep learning and its application in computer vision, such as person re-identification, instance retrieval, and medical image understanding.

Email: liangzheng06@gmail.com     Google Scholar

I am joining Singapore University of Technology and Design as tenure-track Assitant Professor. I am looking for PhD and visiting students with exceptional coding ability and motivation. Drop me an email if you are interested.


  • SIFT Meets CNN: A Decade Survey of Instance Retrieval. TPAMI, Accepted. [PDF] [Bibtex]
  • We have released the code for our paper "Image-Image Domain Adaptation with Preserved Self-Similarity and Domain-Dissimilarity for Person Re-identification". Link
  • We have released the code for our paper "Random Erasing Data Augmentation". Link
  • We have released the code for our paper "Dual-Path Convolutional Image-Text Embedding". Link
  • The state-of-the-art methods for Market-1501+500k [link] and MARS [link] are summarized.
  • We have released ID-level attribute labels for the Market-1501 and Duke datasets. Link
  • Code for our CVPR17 paper "Re-ranking person re-identification with k-reciprocal encoding" is released. We have also released an evaluation protocol for CUHK03 that has one train/test split. Link
  • A large-scale image-based re-ID dataset, Duke, is released based on DukeMTMC. It has a similar scale and evaluation procedure to Market-1501. We have release the train/test split, evaluation code, and benchmarking results. Link
  • Our group in UTS has multiple positions for postdocs, PhDs and interns. Link
  • A person re-identification survey is available. Person Re-identification: Past, Present and Future. Link Baseline Code.
  • An image retrieval survey is available. SIFT Meets CNN: A Decade Survey of Instance Retrieval. Link.
  • We have released the MARS dataset for large scale video based person re-identification. Project Page.
  • A summary of state of the arts on Market-1501 is presented.
  • We have released the PRW (Person Re-identification in the Wild) dataset. See the project page. It supports pedestrian detection and person re-identification.
  • A new person re-identification dataset, the "Market-1501" dataset, is released. See the project page. It has 1501 identities, 32k bounding boxes.
  • Our ArXiv paper "Person Re-identification Meets Image Search" is covered in "MIT Technology Review".
  • We have provided the code for constructing the baseline used in our CVPR'14 papers.
  • We have provided the code for our CVPR'15 paper.
  • I am working/collaborating with

  • Zhedong Zheng, PhD student in UTS. Homepage Github Blog
  • Yutian Lin, PhD student in UTS. Homepage
  • Ruoyu Liu, visiting PhD student in UTS. Homepage
  • Zhun Zhong, PhD student in Xiamen University. Homepage
  • Yuting Hu, PhD student in Tsinghua University. Homepage