Postdoc Researcher at University of Sydney, Australia.
Dr. Weichen Zhang π is a passionate AI Researcher and Engineer specializing in 2D/3D visual applications, currently working as a Postdoctoral Researcher at the University of Sydney π¦πΊ. With a strong academic background and over 8 years of hands-on experience, he has contributed significantly to deep learning, point cloud analysis, and 3D human body reconstruction π€. Formerly the R&D Team Lead at Bodymapp Pty Ltd, he led innovations in AI, SLAM, and AWS MLOps pipelines βοΈ. His work is well-recognized with publications in top-tier venues like CVPR, T-PAMI, and T-IP, and two U.S. patents for AI technologies π.
Professional Profile:
ORCID
Google Scholar
Suitability for Best Researcher Award β Dr. Weichen Zhang
Dr. Weichen Zhang is an outstanding candidate for the Best Researcher Award due to his groundbreaking contributions to the fields of AI, 2D/3D computer vision, and deep learning. With over 8 years of rich academic and industrial experience, he has demonstrated a rare combination of technical depth, publication excellence, and innovation in applied AI. His research has made substantial impact in areas such as point cloud analysis, 3D human body reconstruction, SLAM, and domain adaptation, positioning him as a rising star in AI and computer vision.
π Education
-
π Ph.D. in Deep Transfer Learning, University of Sydney, Australia (2017β2021)
-
π B.I.T. (Hons Class I), University of Sydney, Australia (2013β2016)
-
π High School, Affiliated to Nanjing Normal University, China (2009β2012)
πΌ Experience
-
π§ Postdoctoral Research Associate, University of Sydney (Sep 2024βPresent)
-
π¬ R&D Team Lead & Research Engineer, Bodymapp Pty Ltd (May 2021βAug 2024)
-
π¨βπ« Academic Tutor, University of Sydney (Jul 2019βDec 2019)
π οΈ Professional DevelopmentΒ
Dr. Zhang continually evolves his professional skillset through advanced research collaborations π€, cross-disciplinary project leadership, and active participation in academic publishing π. His professional growth is marked by practical application of machine learning, neural networks, DevOps, and 3D vision technologies across academia and industry. With a strong command of programming languages like Python, C++, and MATLAB, he has built scalable AI systems and AWS MLOps pipelines βοΈ. He also contributes to academic instruction, tutoring deep learning and transfer learning concepts. Zhangβs engagement with top research institutions and conferences keeps him at the cutting edge of AI innovation π.
π§ͺ Research Focus Area
Dr. Weichen Zhangβs research is focused on 2D/3D visual understanding, including point cloud analysis, 3D mesh reconstruction, SLAM, and deep learning model generalization π. His projects span cross-domain recognition, domain adaptation, and privacy-preserving avatar reconstruction technologies π€. He integrates supervised and unsupervised learning methods to solve complex problems in autonomous driving π, human body tracking, and forestry analytics π³. Zhangβs contributions are reflected in CORE A* journals and conferences, showing his commitment to pushing the boundaries of scalable, multimodal, and multi-sensor visual systems in real-world applications π.
π
Awards and Honours
-
π₯ Research Training Program (RTP) Scholarship, University of Sydney (2017β2021)
-
ποΈ Norman I. Price Scholarship, University of Sydney (2018β2019)
-
π CVPR 2018 Paper Spotlight, Salt Lake City, USA (2018)
-
π USYD-CSIRO Summer Research Scholarship, University of Sydney (2016β2017)
Publication Top Notes
π 1. Collaborative and Adversarial Network for Unsupervised Domain Adaptation
-
Authors: W. Zhang, W. Ouyang, W. Li, D. Xu
-
Published in: Proceedings of the IEEE Conference on Computer Vision and Pattern Recognition (CVPR)
-
Year: 2018
-
Cited by: 628
-
Summary: This paper presents a novel framework combining collaborative learning and adversarial training to improve performance in unsupervised domain adaptation. It effectively aligns feature distributions between source and target domains.
π 2. Model Compression using Progressive Channel Pruning
-
Authors: J. Guo, W. Zhang, W. Ouyang, D. Xu
-
Published in: IEEE Transactions on Circuits and Systems for Video Technology
-
Year: 2020
-
Cited by: 74
-
Summary: The authors propose a progressive pruning method that reduces the number of channels in convolutional neural networks, enabling model compression with minimal accuracy loss. The method is data-driven and layer-wise.
π 3. SRDAN: Scale-aware and Range-aware Domain Adaptation Network for Cross-dataset 3D Object Detection
-
Authors: W. Zhang, W. Li, D. Xu
-
Published in: Proceedings of the IEEE Conference on Computer Vision and Pattern Recognition (CVPR)
-
Year: 2021
-
Cited by: 54
-
Summary: SRDAN is a domain adaptation network designed for 3D object detection. It introduces scale-aware and range-aware components to address distribution shifts between datasets, improving detection across varying sensor configurations.
π 4. Progressive Modality Cooperation for Multi-modality Domain Adaptation
-
Authors: W. Zhang, D. Xu, J. Zhang, W. Ouyang
-
Published in: IEEE Transactions on Image Processing, Vol. 30, pp. 3293β3306
-
Year: 2021
-
Cited by: 24
-
Summary: This work proposes a progressive learning scheme where different modalities (e.g., RGB and depth) are learned in stages to enhance domain adaptation. The model progressively integrates information, improving generalization.
π 5. 3D Hand Pose Estimation with Disentangled Cross-Modal Latent Space
-
Authors: J. Gu, Z. Wang, W. Ouyang, W. Zhang, J. Li, L. Zhuo
-
Published in: IEEE Winter Conference on Applications of Computer Vision (WACV), pp. 391β400
-
Year: 2020
-
Cited by: 24
-
Summary: This paper introduces a cross-modal learning framework that disentangles modality-specific and shared features, enabling accurate 3D hand pose estimation from multi-view data.
Conclusion
Dr. Zhangβs sustained contributions to cutting-edge AI research, proven innovation, and recognized leadership in both academia and industry make him an ideal recipient of the Best Researcher Award. His visionary work continues to push the boundaries of what is possible in machine learning and computer vision, with real-world impact and academic distinction.