publications

2026

1. 

   Set-Based Groupwise Registration for Variable-Length, Variable-Contrast Cardiac MRI

   Yi Zhang, Yidong Zhao, Tijmen Toxopeus, and 3 more authors

   In Medical Image Computing and Computer Assisted Intervention (MICCAI), 2026

   PDF Project Page

2025

1. 

   Recurrent Inference Machine for Medical Image Registration

   Yi Zhang, Yidong Zhao, Hui Xue, and 3 more authors

   Medical Image Analysis, 2025

   HTML PDF
2. Sci Rep

   Detecting continuous structural heterogeneity in single molecule localization microscopy data with a point cloud variational auto-encoder

   S. Haghparast, Yi Zhang, Qian Tao, and 2 more authors

   Scientific Reports, 2025
3. arXiv

   Learn2Reg 2024: New Benchmark Datasets Driving Progress on New Challenges

   Lasse Hansen, Wolfgang Heyer, Clara Grossbrohmer, and 4 more authors

   arXiv preprint arXiv:2509.01217, 2025

4. 

   Bridging Classical and Learning-based Iterative Registration through Deep Equilibrium Models

   Yi Zhang, Yidong Zhao, and Qian Tao

   International Conference on Medical Image Computing and Computer Assisted Intervention, 2025

   Abs PDF

   Deformable medical image registration is traditionally formulated as an optimization problem. While classical methods solve this problem iteratively, recent learning-based approaches use recurrent neural networks (RNNs) to mimic this process by unrolling the prediction of deformation fields in a fixed number of steps. However, classical methods typically converge after sufficient iterations, but learning-based unrolling methods lack a theoretical convergence guarantee and show instability empirically. In addition, unrolling methods have a practical bottleneck at training time: GPU memory usage grows linearly with the unrolling steps due to backpropagation through time (BPTT). To address both theoretical and practical challenges, we propose DEQReg, a novel registration framework based on Deep Equilibrium Models (DEQ), which formulates registration as an equilibrium-seeking problem, establishing a natural connection between classical optimization and learning-based unrolling methods. DEQReg maintains constant memory usage, enabling theoretically unlimited iteration steps. Through extensive evaluation on the public brain MRI and lung CT datasets, we show that DEQReg can achieve competitive registration performance, while substantially reducing memory consumption compared to state-of-the-art unrolling methods. We also reveal an intriguing phenomenon: the performance of existing unrolling methods first increases slightly then degrades irreversibly when the inference steps go beyond the training configuration. In contrast, DEQReg achieves stable convergence with its inbuilt equilibrium-seeking mechanism, bridging the gap between classical optimization-based and modern learning-based registration methods.

2024

1. MRM

   Preparation-based B 1+ mapping in the heart using Bloch–Siegert shifts

   Paulina Šiurytė, Joao Tourais, Yi Zhang, and 6 more authors

   Magnetic Resonance in Medicine, 2024

2. 

   PCA-Relax: Deep-learning-based groupwise registration for motion correction of cardiac T_1 mapping

   Yi Zhang, Yidong Zhao, Lu Huang, and 2 more authors

   International Conference on Medical Image Computing and Computer Assisted Intervention, 2024

   Abs PDF

   Quantitative T1 mapping by MRI is an increasingly important tool for clinical assessment of cardiovascular diseases. The cardiac T1 map is derived by fitting a known signal model to a series of baseline images, while the quality of this map can be deteriorated by involuntary respiratory and cardiac motion. To correct motion, a template image is often needed to register all baseline images, but the choice of template is nontrivial, leading to inconsistent performance sensitive to image contrast. In this work, we propose a novel deep-learning-based groupwise registration framework, which omits the need for a template, and registers all baseline images simultaneously. We design two groupwise losses for this registration framework: the first is a linear principal component analysis (PCA) loss that enforces alignment of baseline images irrespective of the intensity variation, and the second is an auxiliary relaxometry loss that enforces adherence of intensity profile to the signal model. We extensively evaluated our method, termed “PCA-Relax”, and other baseline methods on an in-house cardiac MRI dataset including both pre- and post-contrast T1 sequences. All methods were evaluated under three distinct training-and-evaluation strategies, namely, standard, one-shot, and test-time-adaptation. The proposed PCA-Relax showed further improved performance of registration and mapping over well-established baselines. The proposed groupwise framework is generic and can be adapted to applications involving multiple images.

3. 

   TorchT1: GPU-accelerated cardiac T1 mapping with deep learning framework

   Yi Zhang, Yidong Zhao, Yifeng Shao, and 3 more authors

   Medical Imaging with Deep Learning, 2024

   Abs HTML PDF

   Quantitative cardiac T_1 mapping by MRI is an essential non-invasive diagnostic tool for cardiomyopathies. Traditionally, deriving the quantitative T_1 maps of myocardial tissue involves solving non-linear parametric fitting problems per image voxel, which is slow with sequential CPU computation and requires analytical derivation of the Jacobian matrix per signal model. In this paper, we introduce a new paradigm of parametric fitting, termed “TorchT_1", which leverages the powerful parallelization of modern GPUs and well-established functionalities of auto-differentiation in the deep learning framework of PyTorch. TorchT_1 strictly adheres to the signal model and does not require any training. Our method was evaluated on a T_1 mapping dataset with both pre-contrast and post-contrast sequences, and benchmarked by conventional CPU-based fitting and recent end-to-end physics-informed neural network (PINN) mapping. TorchT_1 showed more accurate and reliable mapping quality compared with the pretrained PINN, with a 13-fold acceleration compared with the CPU baseline.

4. MIDL-S

   Anatomy-compliant medical image synthesis by latent diffusion models

   Nuno Miguel Ferreira Capitão, Yidong Zhao, Yi Zhang, and 3 more authors

   Medical Imaging with Deep Learning, 2024
5. SCMR-A

   Myocardium Signal Suppression with t2-preparations for Reduced Physiological Noise in Myocardial ASL

   Masa Bozic-Iven, Yi Zhang, Qian Tao, and 3 more authors

   Journal of Cardiovascular Magnetic Resonance, 2024

2023

1. MICCAIW

   Relaxometry Guided Quantitative Cardiac Magnetic Resonance Image Reconstruction

   Yidong Zhao, Yi Zhang, and Qian Tao

   International Workshop on Statistical Atlases and Computational Models of the Heart, 2023
2. MICCAIW

   Contrast-Agnostic Groupwise Registration by Robust PCA for Quantitative Cardiac MRI

   Xinqi Li, Yi Zhang, Yidong Zhao, and 2 more authors

   International Workshop on Statistical Atlases and Computational Models of the Heart, 2023
3. bioRxiv

   Deep learning-based cell profiling based on neuronal morphology

   Qiang Liu, Francesca J Nicholls, Helen A Rowland, and 8 more authors

   2023

4. 

   The Euclidean Space is Evil: Hyperbolic Attribute Editing for Few-shot Image Generation

   Lingxiao Li, Yi Zhang, and Shuhui Wang

   Int. Conf. Comput. Vis., 2023

   Abs PDF

   Few-shot image generation is a challenging task since it aims to generate diverse new images for an unseen category with only a few images. Existing methods suffer from the trade-off between the quality and diversity of generated images. To tackle this problem, we propose Hyperbolic Attribute Editing(HAE), a simple yet effective method. Unlike prior arts that work in Euclidean space, HAE captures the hierarchy among images using data from seen categories in hyperbolic space. Given a well-trained HAE, images of unseen categories can be generated by moving the latent code of a given image toward any meaningful directions in the Poincaré disk with a fixing radius. Most importantly, the hyperbolic space allows us to control the semantic diversity of the generated images by setting different radii in the disk. Extensive experiments and visualizations demonstrate that HAE is capable of not only generating images with promising quality and diversity using limited data but achieving a highly controllable and interpretable editing process.

2022

1. 

   Clinical Prompt Learning with Frozen Language Models

   Niall Taylor*, Yi Zhang*, Dan Joyce, and 2 more authors

   IEEE Trans. Neural Netw. Learn. Syst., 2022

   Abs PDF

   Prompt learning is a new paradigm in the Natural Language Processing (NLP) field which has shown impressive performance on a number of natural language tasks with common benchmarking text datasets in full, few-shot, and zero-shot train-evaluation setups. Recently, it has even been observed that large but frozen pre-trained language models (PLMs) with prompt learning outperform smaller but fine-tuned models. However, as with many recent NLP trends, the performance of even the largest PLMs such as GPT-3 do not perform well on specialized domains (e.g. medical text), and the common practice to achieve State of the Art (SoTA) results still consists of pre-training and fine-tuning the PLMs on downstream tasks. The reliance on fine-tuning large PLMs is problematic in clinical settings where data is often held in non-GPU environments, and more resource efficient methods of training specialized domain models is crucial. We investigated the viability of prompt learning on clinically meaningful decision tasks and directly compared with more traditional fine-tuning methods. Results are partially in line with the prompt learning literature, with prompt learning able to match or improve on traditional fine-tuning with substantially fewer trainable parameters and requiring less training data. We argue that prompt learning therefore provides lower computational resource costs applicable to clinical settings, that can serve as an alternative to fine-tuning ever increasing in size PLMs.