Tony Ng

PhD @ MatchLab, Imperial College London | ex-Reality Labs, Meta & Scape Technologies | HK, London

about my research

I’m 4-th year Ph.D. student at MatchLab, Imperial College London. I am supervised by Prof. Krystian Mikolajczyk and co-supervised by Dr. Vassileios Balntas.

My research interests are Computer Vision, Deep Learning and 3D Geometry, specifically in combining them to solve the problem of Visual Localization for applications in Augmented Reality. My work include studies on Image Retrieval, Camera Pose Regression, Keypoints Detection & Description. Recently, I have started looking into the problem of privacy & security in computer vision, as well as multi-modal learning using language models (specifically auto-regressive Transformers).

news

Oct 7, 2022 I completed a second research internship at Reality Labs, this time working on multi-modal understanding (text & geometry) using language models.
Jun 24, 2022 Our paper NinjaDesc: Content-Concealing Visual Descriptors via Adversarial Learning was presented at CVPR 2022, New Orleans LA.
Dec 5, 2021 I have been selected as an Outstanding Reviewer (top 8%) at NeurIPS 2021!
Oct 24, 2021 I completed my research internship at Reality Labs, Meta :smiley:
Aug 12, 2020 My first paper, SOLAR: Second-Order Loss and Attention for Image Retrieval is accepted at ECCV 2020! [aXiv][GitHub]

selected publications

  1. CVPR
    NinjaDesc: Content-Concealing Visual Descriptors via Adversarial Learning
    Tony Ng, Hyo Jin Kim, Vincent Lee, Daniel DeTone, Tsun-Yi Yang, Tianwei Shen, Eddy Ilg, Vassileios Balntas, Krystian Mikolajczyk, Chris Sweeney
    CVPR, 2022
  2. arXiv
    Reassessing the Limitations of CNN Methods for Camera Pose Regression
    Tony Ng, Adrian Lopez-Rodriguez, Vassileios Balntas, Krystian Mikolajczyk
    arXiv preprint, 2021
  3. ECCV
    SOLAR: Second-Order Loss and Attention for Image Retrieval
    Tony Ng, Vassileios Balntas, Yurun Tian, Krystian Mikolajczyk
    In ECCV, 2020