About

I specialize in computer vision, machine learning, and image processing. I have worked with Prof. Stephen Gould and Prof. Richard Hartley as a research fellow at the Australian National University. I have received my PhD in computer vision and machine learning under the supervision of Prof. Richard Hartley and Dr. Thalaiyasingam Ajanthan at the Australian National University and Data61, CSIRO. I have also completed my Master of Engineering in signal and image processing under the supervision of Prof. Lei Zhang and Prof. Mengdao Xing at Xidian University and earned my Bachelor of Engineering, with a thesis on watermark for digital color images, under the supervision of Prof. Gui Feng at Huaqiao University.

Publications

Topics: deep declarative networks diffusion models Markov random fields neuron pruning generic computer vision.

Papers and Preprints

  • Joint active and passive beamforming for IRS-assisted monostatic backscatter systems: an unsupervised learning approach
    Sahar Idrees, Salman Durrani, Zhiwei Xu, Xiaolun Jia, and Xiangyun Zhou
    IEEE Transactions on Machine Learning in Communications and Networking, 2024
  • IMPUS: image morphing with perceptually-uniform sampling using diffusion models
    Zhaoyuan Yang, Zhengyang Yu, Zhiwei Xu, Jaskirat Singh, Jing Zhang, Dylan Campbell, Peter Tu, and Richard Hartley
    International Conference on Learning Representations, 2024
  • Adversarial purification with the manifold hypothesis
    Zhaoyuan Yang, Zhiwei Xu, Jing Zhang, Richard Hartley, and Peter Tu
    Association for the Advancement of Artificial Intelligence, 2024
  • Probabilistic and semantic descriptions of image manifolds and their applications
    Peter Tu, Zhaoyuan Yang, Richard Hartley, Zhiwei Xu, Jing Zhang, Yiwei Fu, Dylan Campbell, Jaskirat Singh, and Tianyu Wang
    Frontiers in Computer Science, volume 5, 2023
  • PMaF: deep declarative layers for principal matrix features
    Zhiwei Xu, Hao Wang, Yanbin Liu, and Stephen Gould
    International Conference on Machine Learning Workshop, 2023
  • Towards understanding gradient approximation in equality constrained deep declarative networks
    Stephen Gould, Ming Xu, Zhiwei Xu, and Yanbin Liu
    International Conference on Machine Learning Workshop, 2023
  • Exploiting problem structure in deep declarative networks: two case studies
    Stephen Gould, Dylan Campbell, Itzik Ben-Shabat, Chamin Hewa Koneputugodage, and Zhiwei Xu
    Association for the Advancement of Artificial Intelligence Workshop, 2022
  • ViLESS: least squares on sphere for video classification
    Hao Wang, Zhiwei Xu, and Stephen Gould
    Technical Report, 2021
  • DDDNet-V2: towards unsupervised depth estimation via dual-pixel sensor
    Liyuan Pan, Zhiwei Xu, Liu Liu, Yan Yang, Miaomiao Liu, and Richard Hartley
    Technical Report, 2021
  • RANP: resource aware neuron pruning at initialization for 3D CNNs
    Zhiwei Xu, Thalaiyasingam Ajanthan, Vibhav Vineet, and Richard Hartley
    arXiv:2103.08457 (an extension of 3DV2020 paper), 2021
  • Refining semantic segmentation with superpixel by transparent initialization and sparse encoder
    Zhiwei Xu, Thalaiyasingam Ajanthan, and Richard Hartley
    arXiv:2010.04363, 2020
  • RANP: resource aware neuron pruning at initialization for 3D CNNs
    Zhiwei Xu, Thalaiyasingam Ajanthan, Vibhav Vineet, and Richard Hartley
    International Conference on 3D Vision, pages 180-189, 2020 (Oral)
    [ preprint | code | slides | best student paper ]
  • Fast and differentiable message passing on pairwise Markov random fields
    Zhiwei Xu, Thalaiyasingam Ajanthan, and Richard Hartley
    Asian Conference on Computer Vision, pages 523-540, 2020 (Oral)
  • Interesting components detection for space satellites from ISAR image via feature probabilistic estimation
    Zhiwei Xu, Lei Zhang, Mengdao Xing, Karen M. von Deneen, and Lei Ran
    IET Image Processing, 9(6), pages 506-515, 2015
    [ paper ]
  • Precise cross-range scaling for ISAR images using feature registration
    Zhiwei Xu, Lei Zhang, and Mengdao Xing
    IEEE Geoscience and Remote Sensing Letters, 11(10), pages 1792-1796, 2014
    [ paper ]
  • A fast BP algorithm with wavenumber spectrum fusion for highresolution spotlight SAR imaging
    Lei Zhang, Haolin Li, Zhijun Qiao, and Zhiwei Xu
    IEEE Geoscience and Remote Sensing Letters, 11(9), pages 1460-1464, 2014
    [ paper ]
  • Multiview ISAR image fusion using motion estimation and nonnegative matrix factorization
    Zhiwei Xu, Lei Zhang, Mengdao Xing, and Jia Duan
    Technical Report, 2014
  • Uniform rotational motion compensation for ISAR targets from image domain
    Zhiwei Xu, Lei Zhang, Mengdao Xing, and Gang Xu
    IEEE Conference on Signal and Image Processing (CSIP), pages 166-170, 2014 (Oral)
    [ paper ]
  • Azimuth scaling for ISAR images with feature registration
    Zhiwei Xu, Lei Zhang, and Mengdao Xing
    IEEE Congress on Image and Signal Processing (CISP), pages 1568-1572, 2013 (Oral)
    [ paper ]
  • Component detection using local feature registration
    Zhiwei Xu, Lei Zhang, and Mengdao Xing
    Annual Radar Conference of Xidian University, pages 57-66, 2013 (Paper Proceedings)

Patents

  • CN 2014SR012993, System of ISAR data and images processing V1.0 (software)
    Zhiwei Xu, Lei Zhang, Darong Huang, and Mengdao Xing
  • CN 201310647712.8, Cross-range scaling based on image feature for ISAR images
    Lei Zhang, Zhiwei Xu, Mengdao Xing, Jia Duan, and Gang Xu
  • CN 201310648787.8, Components detection for ISAR targets by local feature registration
    Lei Zhang, Zhiwei Xu, Mengdao Xing, Qi Dong, and Min Wu
  • CN 201410218699.9, Recognition of satellite panel in ISAR images by sum of Gaussian probabilities
    Lei Zhang, Zhiwei Xu, Min Wu, Qi Dong, and Mengdao Xing
  • CN 201410445675.7, ISAR images fusion based on target's characteristics
    Lei Zhang, Zhiwei Xu Xu, and Mengdao Xing
  • CN 201410106245.2, Super-resolution imaging for full Pol-SAR by sectional compressive sensing
    Lei Zhang, Min Wu, Zhiwei Xu, Jia Duan, and Mengdao Xing
  • CN 201310653886.5, Decomposition/reconstruction of ISAR targets by central characteristics model
    Lei Zhang, Qi Dong, Yinghui Quan, Zhiwei Xu, and Gang Xu
  • CN 201310655227.5, Range aligning for ISAR data by linear estimation of polynomials
    Lei Zhang, Qi Dong, Yinghui Quan, Zhiwei Xu, and Gang Xu

Services

  • Guest lecturer for ENGN8503-advanced topics in computer vision, ANU, 2021, 2022, 2023
  • Reviewer for TPAMI, GRSL, CVPR, ECCV, NeurIPS, ACCV, ICML, ANU Computing honours thesis 2021-2023
  • Convenor for ANU-CVML reading group, 2021-2022, and ANU AI+ML+Friends seminar, 2022-2023

Awards

  • Best student paper, international conference on 3D vision, 2020
  • Data61 PhD scholarship, 2017-2021
  • Data61 research project award, 2017-2021
  • ANU HDR fee remission merit scholarship, 2017-2021
  • Outstanding graduate thesis at Xidian University (5 out of 3000 graduate students), 2015
  • Outstanding graduate student at Xidian University (highest honor of postgraduate), 2015
  • National graduate scholarship (top 2% in the school), 2014
  • Outstanding postgraduate at Xidian University (top 5% in the school), 2014
  • Outstanding postgraduate at Xidian University, 2013
  • Outstanding student at Elitist School of Huaqiao University, 2011
  • First-class academic scholarship at Huaqiao University, 2010

Others

  • Programming: PyTorch, Python, MATLAB, CUDA, C/C++, Git, Linux, LaTex, CAD, FPGA, Quartus, FEKO
  • Spare: gym, pets (fishes, a dog, a cat, rabbits, snakes, a lizard, ducks, chickens, geese, birds, tortoises, silkworms)