Z Jane Wang


Electrical and Computer Engineering
The University of British Columbia
2332 Main Mall,Vancouver, BC Canada V6T 1Z1
Office: KAIS 4015
Phone: (604) 822-3229
Email: zjanew_AT_ece.ubc.ca
Z. Jane Wang is a Professor in the Department of Electrical and Computer Engineering at Columbia (UBC), Canada. She received the B.Sc. degree from Tsinghua University in 1996 and the M.Sc. and Ph.D. degrees from the University of Connecticut in 2000 and 2002, respectively, all in electrical engineering. She had been an Research Associate at the University of Maryland, College Park from 2002 to 2004.

Since Aug. 2004, she has been with the ECE dept at UBC.

She is an IEEE Fellow, a Fellow of the Canadian Academy of Engineering (FCAE) and a member of the College of New Scholars, Artists and Scientists of the Royal Society of Canada. Her research interests are in the broad areas of signal/image processing and machine learning, with current focuses on digital media and biomedical data analytics. She has published 180+ journal papers and 120+ peer-reviewed conference papers. She has been key Organizing Committee Member for numerous IEEE conferences and workshops (e.g., the co-Technical Chair for ChinaSIP2014, GlobalSIP2017 and ICIP2021, and the co-General Chair of MMSP2018 and DSLW2021).  She has been Associate Editor for the IEEE TSP, SPL, TMM, TIFS, TBME, and SPM, and Area Editor of SPM, and is currently serving as Editor-in-Chief for IEEE SPL.


For a complete list of publications, see my Google Scholar page.

Research Interests

Signal/image processing, machine learning/deep learning, biomedical data analysis, digital media security

Selected Research Projects

Deep learning interpretability: We focus on visual interpretation for deep convolutional neural networks (DCNNs) in image analysis, mainly in image classification. Interpretation of their internal network mechanism has been increasingly important, while the network decision making logic is still an open issue. We first give the interpretation of DCNNs from 3 levels (network layer level, network channel level, and relationship between layers). Further, to have a more comprehensive interpretation, we propose the CHAIN interpretation to explain the DCNN network bottom-up decision-making process in a hierarchical inference way.
Deep learning with Limited Annotation Budget (transfer learning, partial annotation): Fully annotating a largescale dataset is an arduous and expensive task. Also, for some specific applications (e.g., medical problems), annotations could be ambiguous or unavailable. Our goal is to make use of the limited annotation information in a more efficient way. We mainly focus on the following two strategies of efficiently utilizing the limited annotations: unsupervised transfer learning and partial annotation learning.
3D Human Pose Estimation from Monocular Camera: Our goal is to train a deep learning network that outputs 3D human keypoints given a monocular image. We develop advanced weak-supervision frameworks to improve the performance of reconstructing 3D human poses from 2D images in real applications (e.g., Parkinson Disease patient monitoring).
Adversarial Deep Learning in Digital Image Forensics: This gives a brief review of my group's research efforts in the intersection of deep learning and digital media security -- adversarial deep learning, including both adversarial attacks and adversarial defenses. Deep learning has achieved state-of-the-art performances in many applications. Unfortunately, both digital images and current deep learning models are vulnerable to manipulations and attacks, giving rise to security, privacy and reliability issues in practical applications.
Deep learning for medical image analysis: During the last few years, deep learning has rapidly become popular for medical image analysis, including classification, object detection, segmentation, registration, and other tasks. A few example studies in our group include skin image analysis, 2D/3D medical image registration, image retrieval and US-based cancer classification.
Neural connectivity modeling: We have made pioneering contributions to fMRI brain connectivity modeling. Our recent group members continued on the brain connectivity modeling and cortico-muscular Interactivity directions, and extended the research into noninvasive treatment of Parkinson’s disease using Galvanic Vestibular Stimulation (GVS).
Back to Faculty
Back to ECE