Selected work

Representative Papers

Low performing pixel correction in computed tomography with unrolled network and synthetic data training

Jan 28, 2026

This work addresses the challenge of ring and streak artifacts in computed tomography (CT) caused by low-performance pixels (LPP) in detectors, which severely compromise clinical diagnosis. The authors propose a dual-domain joint correction method based on an unrolled network that, for the first time, integrates the CT geometric forward model into a deep learning framework to collaboratively model LPP-induced artifacts in both the sinogram and image domains. Training data are synthesized from natural images, eliminating the need for real clinical defect measurements and enabling end-to-end training as well as cross-scanner deployment. Under simulated detector defect rates of 1–2%, the proposed method significantly outperforms existing techniques, effectively suppressing artifacts while demonstrating strong practicality and generalization capability.

1 citationsRead paper

Unsupervised Adaptation of 3D CT Foundation Models for 3D CBCT Segmentation

Aug 27, 2026

本文提出一种无监督领域自适应框架，通过减少冗余特征对齐来解决3D CBCT分割中数据标注稀缺和模态差异问题。

0 citationsRead paper

A Unified DINOv2-Based Framework for LVEF Estimation, GLS Dysfunction Classification, and Early Cardiotoxicity Prediction

Aug 13, 2026

This study addresses the challenge of cardiac function assessment without cardiac cycle segmentation in cardio-oncology by proposing a unified framework based on DINOv2. By integrating LoRA fine-tuning, temporal aggregation, and physiology-guided hybrid regression, the method innovatively enables cycle-free inference and multi-task joint learning. This approach validates the potential of visual foundation models for medical time-series analysis while significantly reducing reliance on annotated data. Experimental results demonstrate superior performance in left ventricular ejection fraction estimation (MAE 4.64%), global longitudinal strain classification (AUC 76.48%), and early cardiotoxicity prediction (AUC 70.26%). Collectively, this work establishes an efficient paradigm for precise clinical assessment in cardio-oncology, highlighting the efficacy of adapting vision foundation models to complex physiological signal interpretation without extensive manual annotation.

0 citationsRead paper

CUPA-T2*: Covariance-Aware Uncertainty Propagation and Alignment for T2* Mapping in Accelerated MRI

Aug 09, 2026

This work addresses the degradation in T2* quantification accuracy in accelerated MRI caused by artifacts and noise from undersampled reconstruction, compounded by the lack of effective modeling of uncertainty propagation to downstream fitting. The study introduces the first explicit covariance-aware uncertainty propagation framework from reconstruction to T2* fitting in accelerated T2* mapping. It estimates voxel-wise multi-echo reconstruction uncertainty via Monte Carlo Dropout, propagates this uncertainty to T2* fitting using a covariance-aware sampling strategy, and aligns predicted variances with reconstruction uncertainties through a heteroscedastic MLP and a correlation-based regularization term. Experiments demonstrate that the proposed method significantly improves T2* fitting performance in white matter under high acceleration in brain MRI, enhances uncertainty consistency, and yields interpretable voxel-level uncertainty maps.

0 citationsRead paper

Few-Shot Concept Prompt Learning for Segmentation Foundation Models via Visual Grounding

Aug 02, 2026

This work addresses the performance limitations of existing foundation models for medical image segmentation, which rely on natural language prompts but suffer from a scarcity of image–text paired data. The authors propose FS-CPL, a novel approach that introduces visually grounded continuous concept prompts to replace conventional textual prompts. Operating with a frozen backbone, FS-CPL requires only a few image–mask samples and optimizes learnable prompt embeddings through mask-supervised learning, eliminating the need for additional text data or backbone retraining. The method achieves substantial performance gains, improving Dice scores by up to 0.62 across four benchmarks—BUSI, HC18, TN3K, and CVC-Clinic—and demonstrates compatibility with both SAM3 and Medical SAM3, confirming its effectiveness and generalizability.

0 citationsRead paper

Recent publications

Latest Papers

Unsupervised Adaptation of 3D CT Foundation Models for 3D CBCT Segmentation

Aug 27, 2026

本文提出一种无监督领域自适应框架，通过减少冗余特征对齐来解决3D CBCT分割中数据标注稀缺和模态差异问题。

0 citationsRead paper

A Unified DINOv2-Based Framework for LVEF Estimation, GLS Dysfunction Classification, and Early Cardiotoxicity Prediction

Aug 13, 2026

This study addresses the challenge of cardiac function assessment without cardiac cycle segmentation in cardio-oncology by proposing a unified framework based on DINOv2. By integrating LoRA fine-tuning, temporal aggregation, and physiology-guided hybrid regression, the method innovatively enables cycle-free inference and multi-task joint learning. This approach validates the potential of visual foundation models for medical time-series analysis while significantly reducing reliance on annotated data. Experimental results demonstrate superior performance in left ventricular ejection fraction estimation (MAE 4.64%), global longitudinal strain classification (AUC 76.48%), and early cardiotoxicity prediction (AUC 70.26%). Collectively, this work establishes an efficient paradigm for precise clinical assessment in cardio-oncology, highlighting the efficacy of adapting vision foundation models to complex physiological signal interpretation without extensive manual annotation.

0 citationsRead paper

CUPA-T2*: Covariance-Aware Uncertainty Propagation and Alignment for T2* Mapping in Accelerated MRI

Aug 09, 2026

This work addresses the degradation in T2* quantification accuracy in accelerated MRI caused by artifacts and noise from undersampled reconstruction, compounded by the lack of effective modeling of uncertainty propagation to downstream fitting. The study introduces the first explicit covariance-aware uncertainty propagation framework from reconstruction to T2* fitting in accelerated T2* mapping. It estimates voxel-wise multi-echo reconstruction uncertainty via Monte Carlo Dropout, propagates this uncertainty to T2* fitting using a covariance-aware sampling strategy, and aligns predicted variances with reconstruction uncertainties through a heteroscedastic MLP and a correlation-based regularization term. Experiments demonstrate that the proposed method significantly improves T2* fitting performance in white matter under high acceleration in brain MRI, enhances uncertainty consistency, and yields interpretable voxel-level uncertainty maps.

0 citationsRead paper

Few-Shot Concept Prompt Learning for Segmentation Foundation Models via Visual Grounding

Aug 02, 2026

This work addresses the performance limitations of existing foundation models for medical image segmentation, which rely on natural language prompts but suffer from a scarcity of image–text paired data. The authors propose FS-CPL, a novel approach that introduces visually grounded continuous concept prompts to replace conventional textual prompts. Operating with a frozen backbone, FS-CPL requires only a few image–mask samples and optimizes learnable prompt embeddings through mask-supervised learning, eliminating the need for additional text data or backbone retraining. The method achieves substantial performance gains, improving Dice scores by up to 0.62 across four benchmarks—BUSI, HC18, TN3K, and CVC-Clinic—and demonstrates compatibility with both SAM3 and Medical SAM3, confirming its effectiveness and generalizability.

0 citationsRead paper

Geospatial Diffusion-based Evolution Synthesis (GeoDES) for Storm-Centered Weather Augmentation

Jul 21, 2026

This study addresses the longstanding challenge in numerical weather prediction wherein existing models struggle to simultaneously achieve high resolution and physical consistency in representing large-scale storm structures—regional models suffer from sparse observational data, while global models are computationally prohibitive and lack sufficient resolution. To overcome this, the authors propose GeoDES, a storm-centric image-to-video diffusion model that introduces, for the first time in meteorological synthesis, a tailored diffusion mechanism incorporating geospatial constraints and storm evolution priors. Evaluated on a North Atlantic test set, GeoDES reduces peak vorticity error by 52% and improves anomaly correlation by 8% over current state-of-the-art methods, demonstrating superior fidelity and physical plausibility. This approach establishes a new paradigm for data augmentation and stress-testing of forecasting systems.

0 citationsRead paper