Yixin Chen 陈亦新 PhD candidate in Medical Imaging Technology, Peking University
Principal investigator of 1 National Natural Science Foundation of China (NSFC) project. I develop universal segmentation and cross-modality adaptation methods for PET, CT and MRI, and use whole-body PET/CT to study multi-organ metabolic alterations. My first-author papers appear in Nature Communications , IEEE TMI , IEEE TRPMS , CVPR 2026 Findings and MICCAI 2024 (poster), with a further paper accepted in EJNMMI .
Representative Works
Nature Communications, 2025
Modality-projection universal model for comprehensive full-body medical imaging segmentation
With Shandong Provincial Qianfoshan Hospital, MPUM maps shared anatomy to modality-specific kernels for whole-body PET, CT and MRI segmentation.
Chen et al., Nature Communications , CC BY-NC-ND 4.0
IEEE TMI, 2024
PCNet: Prior Category Network for CT Universal Segmentation Model
PCNet encodes clinical relationships among anatomical categories as prompts, a hierarchy and a consistency constraint to guide universal CT segmentation.
Original diagram from the PCNet repository
EJNMMI, accepted 2026
Whole-Body [18F]FDG PET/CT Characterizes Multi-Organ Metabolic Alterations in MASLD
With Henan Medical University First Affiliated Hospital (Xinxiang) and Shandong Provincial Qianfoshan Hospital, a two-centre study of 294 adults identified and replicated 16 MASLD-related imaging features.
Chen, Zhou, Guo et al.
2026 Whole-Body [18 F]FDG PET/CT Characterizes Multi-Organ Metabolic Alterations in MASLD EJNMMI , 2026Accepted Authors' manuscript Whole-body FDG PET/CT is used to examine metabolic changes across organs in MASLD.
Yixin Chen , Chuanzhi Zhou, Lijuan Guo, Huiling Duan, Fengjuan Li, Lihong Yang, Yanhua Duan, Hui Li, Ruoyan Xu, Zhaoping Cheng, Zhaoheng Xie
NeuronCtrl: Geometry-Aware Safe Closed-Loop Generative Control for Neuronal Microenvironment Dynamics ICML , 2026Spotlight Figure source NeuronCtrl combines a generative model with closed-loop control to guide neuronal microenvironment dynamics.
Haowei Xu, Yixin Chen , Wanyi Fu, Hongbin Han, Zhaoheng Xie
Neural Network Approximation Aided Kinetic Parameter Inference of Compartment Models in Dynamic PET IEEE TRPMS , 2026A neural approximation of compartment-model equations is used to infer kinetic parameters from dynamic PET.
Wenrui Shao, Yixin Chen , Ourui Fu, Tian Li, Zhi Yang, Nan Li, Hongbin Han, Xiangxi Meng, Zhaoheng Xie
PHATE-Net: Differentiable Pseudotime Learning for Trustworthy Disease Trajectories in PET CVPR Findings , 2026Figure source PHATE-Net learns a disease progression timeline from PET images by making pseudotime estimation differentiable.
Yixin Chen , Yan Wang, Wenrui Shao, Zhaoheng Xie
Diagnosis-Conditioned Retrieval for Structured Pathology Workflows MICCAI REG² challenge workshop , 2026Poster Predicted organ and diagnosis labels guide retrieval of a structured pathology workflow for each whole-slide image.
Yixin Chen
Acquisition-Structured nnU-Net Ensembles for Multi-Sequence Cardiac MRI Segmentation and Quantification MICCAI CMRSeg workshop , 2026Selected for oral presentation Sequence-specific nnU-Net ensembles segment cardiac MRI and derive ejection fraction and scar mass.
Yixin Chen , Yizhen Peng
Failure-Mode-Driven Multimodal Mitral Valve Segmentation for the MVAA 2026 Challenge MICCAI MVAA challenge workshop , 2026Poster Modality-specific ensembles segment the mitral valve in cardiac CT, 3D ultrasound and surgical video.
Yixin Chen
2025 Modality-projection universal model for comprehensive full-body medical imaging segmentation Nature Communications , 2025Figure source , CC BY-NC-ND 4.0 A modality-projection model segments anatomical structures across whole-body medical images.
Yixin Chen * , Lin Gao* , Yajuan Gao, Rui Wang, Jingge Lian, Xiangxi Meng, Yanhua Duan, Leiying Chai, Hongbin Han, Zhaoping Cheng, Zhaoheng Xie
VP-SFDA: Visual Prompt Source-Free Domain Adaptation for Cross-Modal Medical Image Health Data Science , 2025Figure source Visual prompts help adapt a segmentation model to a new imaging modality without access to the source data.
Yixin Chen , Yan Wang, Zhaoheng Xie
Neural network-aided unsupervised input function estimation for dual-time-window PET Patlak analysis EJNMMI Physics , 2025Figure source A neural network estimates the missing input function needed for Patlak analysis from dual-time-window PET scans.
Wenrui Shao, Yarong Zhang, Fen Du, Fangxiao Cheng, Yixin Chen , Xiangxi Meng, Ying Liang, Zhaoheng Xie
A Physiology-Supervised Neural Operator Framework for Kinetic Parameter Estimation in Dynamic PET IEEE NSS/MIC/RTSD , 2025A physiology-supervised neural operator estimates kinetic parameters from dynamic PET data.
W. Shao, Yixin Chen , Z. Wang, R. Xu, X. Meng, Z. Xie
2024 PCNet: Prior Category Network for CT Universal Segmentation Model IEEE TMI , 2024Figure source PCNet uses prior anatomical categories to guide universal segmentation of CT structures.
Yixin Chen , Yajuan Gao, Lei Zhu, Wenrui Shao, Yanye Lu, Hongbin Han, Zhaoheng Xie
LUCIDA: Low-Dose Universal-Tissue CT Image Domain Adaptation for Medical Segmentation MICCAI , 2024Poster Figure source LUCIDA uses segmentation reconstruction and Fourier adaptation without low-dose CT labels; its ensemble recognizes 112 tissue types.
Yixin Chen , Xiangxi Meng, Yan Wang, Shuang Zeng, Xi Liu, Zhaoheng Xie
Structure-Enhanced Unsupervised Domain Adaptation for CT Whole-Brain Segmentation IEEE TRPMS , 2024Structure-aware domain adaptation is applied to whole-brain CT segmentation.
Yixin Chen * , Yajun Gao* , Lei Zhu, Jianan Li, Yan Wang, Jiakui Hu, Hongbin Han, Yanye Lu, Zhaoheng Xie
Neural Network-Asissted Dual-time-window Acquisition: Modeling Incomplete Input Functions for Patlak Analysis IEEE NSS/MIC/RTSD , 2024A neural model reconstructs the incomplete input function in dual-time-window PET for Patlak analysis.
W. Shao, Y. Zhang, Yixin Chen , X. Meng, Y. Liang, Z. Xie
2023 FVP: Fourier Visual Prompting for Source-Free Unsupervised Domain Adaptation of Medical Image Segmentation IEEE TMI , 2023Figure source Fourier-domain visual prompts adapt medical image segmentation when source-domain data are unavailable.
Yan Wang, Jian Cheng, Yixin Chen , Shuai Shao, Lanyun Zhu, Zhenzhou Wu, Tao Liu, Haogang Zhu
Rethinking Disentanglement in Unsupervised Domain Adaptation for Medical Image Segmentation IEEE EMBC , 2023Figure source The paper revisits feature disentanglement for unsupervised domain adaptation in medical segmentation.
Yan Wang* , Yixin Chen * , Yingying Zhang, Haogang Zhu
Echocardiography-based AI for detection and quantification of atrial septal defect Frontiers in Cardiovascular Medicine , 2023Figure source Echocardiography-based AI detects atrial septal defects and quantifies their size.
Xixiang Lin* , Feifei Yang* , Yixin Chen * , Xu Chen, Wenjun Wang, Wenxiu Li, Qiushuang Wang, Liwei Zhang, Xin Li, Yujiao Deng, Haitao Pu, Xiaotian Chen, Xiao Wang, Dong Luo, Peifang Zhang, Daniel Burkhoff, Kunlun He
2022 MSGAN: Multi-Stage Generative Adversarial Networks for Cross-Modality Domain Adaptation IEEE EMBC , 2022A multi-stage generative adversarial network addresses cross-modality domain adaptation for medical images.
Yan Wang* , Yixin Chen * , Wenjun Wang, Haogang Zhu
A deep learning model system for diagnosis and management of adnexal masses Cancers , 2022Figure source A deep-learning system analyzes ultrasound images to support the diagnosis and management of adnexal masses.
Jianan Li* , Yixin Chen * , Minyu Zhang, Peifang Zhang, Kunlun He, Fengqin Yan, Jingbo Li, Hong Xu, Daniel Burkhoff, Yukun Luo, Longxia Wang, Qiuyang Li
Echocardiography-based AI detection of regional wall motion abnormalities and quantification of cardiac function in myocardial infarction Frontiers in Cardiovascular Medicine , 2022Figure source An echocardiography-based AI system detects regional wall-motion abnormalities and measures cardiac function after myocardial infarction.
Xixiang Lin, Feifei Yang, Yixin Chen , Xiaotian Chen, Wenjun Wang, Xu Chen, Qiushuang Wang, Liwei Zhang, Huayuan Guo, Bohan Liu, Liheng Yu, Haitao Pu, Peifang Zhang, Zhenzhou Wu, Xin Li, Daniel Burkhoff, Kunlun He
* Equal contribution.
Challenges and presentations 2026 autoPET V: second place in Category 1CYX-AI combined neural-network corrections with local threshold updates for interactive PET/CT lesion segmentation. Oral presentation at the MICCAI 2026 autoPET challenge workshop.
2026 CMR-MULTI: second placeMulti-sequence cardiac MRI segmentation. Method paper selected for a CMRSeg challenge oral presentation.
Photo to be added
2026 REG²: fourth placeThe REG² challenge generates structured pathology workflows from whole-slide images. Method paper selected for a MICCAI 2026 challenge workshop poster.
2026 CHIMERA-agent: Innovative Agent Solution AwardTeam PKU Intelligence, MICCAI 2026.
News coverage 2026-10-10 Institute of Medical Technology, Peking University Health Science Center
2026-01-13 Institute of Medical Technology, Peking University Health Science Center
Teaching, patents and reviewing
Patents 12 public patent records: 5 granted Chinese invention patents, 6 published Chinese invention applications, and 1 published PCT international application.
Nature Communications Early Career Co-Reviewer
Teaching assistant, Peking University Principles of Medical Imaging Equipment
Computer vision course instructor Independently recorded introductory and intermediate courses in computer vision, covering machine learning fundamentals and computer vision algorithms.
“机器学习炼丹术” on WeChat Shared notes on machine-learning practice and medical imaging through my technical WeChat account. Visitors around the world Visits since October 2026
Loading visits…
Approximate location from IP. Only country or region totals are stored.
Map: MapSVG / SVG Maps , CC BY 4.0
