Projects/ code/ data

NEW!!! ★ S Gao, H Zhou, Y Gao & X Zhuang. Joint Modeling of Image and Label Statistics for Enhancing Model Generalizability of Medical Image Segmentation. MICCAI 2022 paper code
★ Z Gao, L Li, F Wu, S Wang & X Zhuang. Decoupling Predictions in Distributed Learning for Multi-Center Left Atrial MRI Segmentation. MICCAI 2022 code
★ K Zhang & X Zhuang. ShapePU: A New PU Learning Framework Regularized by Global Consistency for Scribble Supervised Cardiac Segmentation. MICCAI 2022 code&data
★ K Zhang & X Zhuang. CycleMix: A Holistic Strategy for Medical Image Segmentation from Scribble Supervision. CVPR 2022 code&data
★ S Gao & X Zhuang. Bayesian Image Super-Resolution with Deep Modeling of Image Statistics. IEEE T PAMI 2022+ link code
★ W Ding, L Li, X Zhuang* & L Huang*: Cross-Modality Multi-Atlas Segmentation via Deep Registration and Label Fusion. IEEE JBHI 2022+ code

Publications with resources We investigate algorithms for multi-modality image analysis, particularly for combined computing, registration:
★ Xiahai Zhuang: Multivariate mixture model for myocardial segmentation combining multi-source images. IEEE T PAMI 41 (12), 2933-2946, 2019. code note&demo
★ X Luo & X Zhuang. MvMM-RegNet: A new image registration framework based on multivariate mixture model and neural network estimation. In Proc. MICCAI 2020. code
★ W Ding, L Li, X Zhuang* & L Huang*: Cross-Modality Multi-Atlas Segmentation via Deep Registration and Label Fusion. IEEE JBHI 2022+ code

We investigate image restoration (super-resolution, denoise and deblur) and reconstruction:
★ S Gao & X Zhuang. Bayesian Image Super-Resolution with Deep Modeling of Image Statistics. IEEE T PAMI 2022+ link code
★ S Gao & X Zhuang. Rank-One Network: An Effective Framework for Image Restoration. IEEE T PAMI 44(6), 3224 - 3238, 2022. code
★ S Gao & X Zhuang. Multi-scale deep neural networks for real image super-resolution. CVPR Workshop 2019. code

We investigate the topic of cardiac pathology analysis, for automatic left atrial and atrial scar segementation and quantification:
☆ L Li, V.A Zimmer, J.A Schnabel, X Zhuang. Medical Image Analysis on Left Atrial LGE MRI for Atrial Fibrillation Studies: A Review, Medical Image Analysis, 102360, 2022.
☆ L Li, V.A Zimmer, J.A Schnabel, X Zhuang. AtrialJSQnet: A New Framework for Joint Segmentation and Quantification of Left Atrium and Scars Incorporating Spatial and Shape Information. Medical Image Analysis, 76, 102303, 2022. code
☆ L Li, F Wu, G Yang, L Xu, T Wong, R Mohiaddin, D Firmin, J Keegan, X Zhuang. Atrial Scar Quantification via Multi-Scale CNN in the Graph-Cuts Framework. Medical Image Analysis 2020. code
☆ L Li, V.A Zimmer, J.A Schnabel, X Zhuang. AtrialGeneral: Domain Generalization for Left Atrial Segmentation of Multi-Center LGE MRIs. MICCAI 2021.

We investigate weak/unsupervised segmentation, domain adpatation and domain generalization, particularly for cross-modality image segmentation:
★ K Zhang & X Zhuang. CycleMix: A Holistic Strategy for Medical Image Segmentation from Scribble Supervision. CVPR 2022 code&data
★ C Pei, F Wu, L Huang, X Zhuang. Disentangle domain features for cross-modality cardiac image segmentation. Medical Image Analysis 2021 code
★ F Wu & X Zhuang. Unsupervised Domain Adaptation with Variational Approximation for Cardiac Segmentation. IEEE TMI 2021. code&data
★ F Wu & X Zhuang. CF Distance: A new domain discrepancy metric and application to explicit domain adaptation for cross-modality cardiac image segmentation. IEEE TMI 2020. code and data (2D WHS)

The topic of medical image computing/analysis:
☆ K Zhang & X Zhuang: CMRadjustNet: Recognition and standardization of cardiac MRI orientation via multi-tasking learning and deep neural networks, MyoPS2020 (a tool for detecting and changing orientation of 2D cardiac bSSFP/LGE/T2 MRI) code
☆ Q Yue, X Luo, Q Ye, L, Xu, X Zhuang. Cardiac Segmentation from LGE MRI Using Deep Neural Network Incorporating Shape and Spatial Priors. MICCAI 2019, LNCS 11765, pp. 559-567, 2019. code
☆ Y Huang, J Xu, Y Zhou, T Tong, X Zhuang. Diagnosis of Alzheimer’s Disease via Multi-Modality 3D Convolutional Neural Network. Frontiers in Neuroscience 2019. code

LAScarQS'22 Challenge

We will provide 200 subjects, and the target of this challenge is to quantify or segment the myocardial pathology (scars) of left atrial walls from LGE MRI of patients suffering atrial fibrillation.

MyoPS'20 Challenge

The challenge provides aligned cardiac T2/bSSFP/LGE MRI, which had been manually segmented the scars, edema, normal myocardium and LV/RV blood pool, from 45 patients. The target is to segment the myocardial pathology combining multi-sequence CMR. Specifically, the myocardium will be classified into normal, infarcted and edema regions, which is important for the diagnosis and treatment management of patients.

MS-CMRSeg'19 Challenge

We provide 45 multi-sequence CMR images from patients who underwent cardiomyopathy. Each patient had been scanned using the three CMR sequences, i.e. the LGE, T2 and bSSFP. The task of this challenge is to segment the ventricles and myocardium from LGE CMR, combing with other two sequences (T2 and bSSFP) from same patients, which can be used to assist the LGE CMR segmentation.

MMWHS Challenge

The challenge provides 120 multi-modality cardiac images acquired in real clinical environment. It aims at creating an open and fair competition for various research groups to test and validate their methods, particularly for the multi-modality whole heart segmentation. It is not only to benchmark various whole heart segmentation algorithms, but also to cover the topic of general cardiac image segmentation and registration and modeling.

S/W zxhproj

zxhproj is a medical image computing platform, being developed and maintained by Xiahai Zhuang since 2004. Based on it, several image registration and segmentation tools have been developed, such as multivariate mixture model, cardiac segmentation, locally affine registration method (LARM), spatially encoded mutual information (SEMI), as well as other image/vector field processing tools. The tools work with nifty (nii/nii.gz) and gipl short int data formats.