Ziyang Wang

Ziyang Wang
Room
007,
Wolfson Building,
Parks Road, Oxford OX1 3QD
United Kingdom
Biography
I have been a DPhil student in Computer Science and St-Hilda's College since 2019. Prior to that, I completed an MRes degree with distinction at Imperial College London in 2018, and a BEng degree in Automation at Xi'an Jiaotong University, China in 2017.
My current research involves techniques for deep learning and computer vision, with practical application in medical image segmentation. Some of my other interests are around image registration, depth estimation, stereo matching, gait analysis, and robotics.
I study image segmentation machine learning models around U-Net, LinkNet, PSPNet, FPN, Atrous CNN, Pooling layer design, Attention, label editing, with supervised learning, weakly-supervised learning, semi-supervised learning. My research improves their performance on Ultrasound, MRI and CT medical data.
I have served as a TA in Information Theory at the Mathematical Institute.
Selected Publications
-
Weakly Supervised Medical Image Segmentation through Dense Combinations of Dense Pseudo−Labels
Ziyang Wang and Irina Voiculescu
In International Conference on Medical Image Computing and Computer Assisted Intervention (MICCAI). Data Engineering in Medical Imaging (DEMI) Workshop. October, 2023.
Best Paper Award‚ Data Engineering in Medical Imaging (DEMI)
Details about Weakly Supervised Medical Image Segmentation through Dense Combinations of Dense Pseudo−Labels | BibTeX data for Weakly Supervised Medical Image Segmentation through Dense Combinations of Dense Pseudo−Labels | Link to Weakly Supervised Medical Image Segmentation through Dense Combinations of Dense Pseudo−Labels
-
Exigent Examiner and Mean Teacher: A Novel 3D CNN−based Semi−Supervised Learning Framework for Brain Tumor Segmentation
Ziyang Wang and Irina Voiculescu
In International Conference on Medical Image Computing and Computer Assisted Intervention (MICCAI). The 2nd Workshop of Medical Image Learning with Limited & Noisy Data (MILLanD). October, 2023.
Details about Exigent Examiner and Mean Teacher: A Novel 3D CNN−based Semi−Supervised Learning Framework for Brain Tumor Segmentation | BibTeX data for Exigent Examiner and Mean Teacher: A Novel 3D CNN−based Semi−Supervised Learning Framework for Brain Tumor Segmentation | Link to Exigent Examiner and Mean Teacher: A Novel 3D CNN−based Semi−Supervised Learning Framework for Brain Tumor Segmentation
-
Dealing with Unreliable Annotations: A Noise−Robust Network for Semantic Segmentation through A Transformer−Improved Encoder and Convolution Decoder
Ziyang Wang and Irina Voiculescu
In Applied Sciences. Vol. 13. No. 13. Pages 7966. July, 2023.
Details about Dealing with Unreliable Annotations: A Noise−Robust Network for Semantic Segmentation through A Transformer−Improved Encoder and Convolution Decoder | BibTeX data for Dealing with Unreliable Annotations: A Noise−Robust Network for Semantic Segmentation through A Transformer−Improved Encoder and Convolution Decoder | Link to Dealing with Unreliable Annotations: A Noise−Robust Network for Semantic Segmentation through A Transformer−Improved Encoder and Convolution Decoder