Publications

JMed48k: A Multi-Profession Japanese Medical Licensing Benchmark for Vision-Language Model Evaluation

Figure 1 · Benchmark overview

Revitalizing Medical Time Series with Vision-Informed Retrieval: A Vision-Language Perspective

Figure 2 · ViRe framework

SphereFlow: Missing Modality Imputation via Geometric Transport on Hypersphere

Figure 3 · SphereFlow framework

Think Before You Move: Latent Motion Reasoning for Text-to-Motion Generation

Figure 4 · Latent Motion Reasoning framework

JMedEthicBench: A Multi-Turn Adversarial Benchmark for Japanese Medical Ethics Alignment in LLMs

Figure 1 · Benchmark construction pipeline

Taming arbitrary modality missingness and imbalance: A unified graph-MoE framework for Alzheimer’s disease diagnosis

MammoDG: Generalisable Deep Learning Advances Cross-Domain Multi-Center Breast Cancer Screening

DenseControl: Instance-Level Controllable Synthesis of Dense Crowd Image

Beyond Time Shifts: Adapting Omni-LLM as a Reference-Free Evaluator for Generative Audio-Visual Models

Progression as Latent Drift: Generative Forecasting of Slow-Evolving Pathologies

BrainAnytime: Anatomy-Aware Cross-Modal Pretraining for Brain Image Analysis with Arbitrary Modality Availability

PRA-PoE: Robust Multimodal Alzheimer’s Disease Classification under Arbitrary Modality Missingness

Evo-RAD: Navigating Rare Retinal Disease Diagnosis via Self-Evolving Agentic Retrieval

Guided by the Plan: Enhancing Faithful Autoregressive Text-to-Audio Generation with Guided Decoding

Decentralized Attention Fails Centralized Signals: Rethinking Transformers for Medical Time Series

Bridging Radiology and Pathology Foundation Models via Concept-Based Multimodal Co-Adaptation

Moving Beyond Functional Connectivity: Time-Series Modeling for fMRI-Based Brain Disorder Classification

Exploring Scale Shift in Crowd Localization under the Context of Domain Generalization

MorphoNet: Morphological sub-region-based structure learning for WSI analysis

Predicting Diabetic Macular Edema Treatment Responses Using OCT: Dataset and Methods of APTOS Competition

Figure 2 · APTOS competition workflow

AD-CARE: A Guideline-grounded, Modality-agnostic LLM Agent for Real-world Alzheimer's Disease Diagnosis with Multi-cohort Assessment, Fairness Analysis, and Reader Study

Project Imaging-X: A Survey of 1000+ Open-Access Medical Imaging Datasets for Foundation Model Development

Knowledge-guided adaptation of pathology foundation models effectively improves cross-domain generalization and demographic fairness

Language model based text-to-audio generation: Anti-causally aligned collaborative residual transformers

ADAgent: Llm agent for alzheimer’s disease analysis with collaborative coordinator

HIBMatch: Hypergraph Information Bottleneck for Semi-Supervised Alzheimer's Progression

Ada-FCN: Adaptive Frequency-Coupled Network for fMRI-Based Brain Disorder Classification

Model-agnostic personalized adaptation for segment anything model

Medical Knowledge Intervention Prompt Tuning for Medical Image Classification

ADFound: A Foundation Model for Diagnosis and Prognosis of Alzheimer's Disease

Latent Imputation before Prediction: A New Computational Paradigm for De Novo Peptide Sequencing

AOR: Anatomical Ontology-Guided Reasoning for Medical Large Multimodal Model in Chest X-Ray Interpretation

A buoyant plasmonic microbubble-based SERS sensing platform for amyloid-beta protein detection in Alzheimer’s disease

Figure 1: SERS substrate preparation and protein sensing

MMR-Mamba: Multi-modal MRI reconstruction with Mamba and spatial-frequency information fusion

Figure 1: MMR-Mamba framework

Toward Lung Ultrasound Automation: Fully Autonomous Robotic Longitudinal and Transverse Scans Along Intercostal Spaces

Figure 2: Robotic lung ultrasound system

Older Fallers’ Comprehensive Neuromuscular and Kinematic Alterations in Reactive Balance Control: Indicators of Balance Decline or Compensation? A Pilot Study

Figure 2: Neuromuscular signal analysis

Synchronized Video-to-Audio Generation via Mel Quantization-Continuum Decomposition

Figure 2: Mel-QCD generation pipeline

Tau-targeting multifunctional nanocomposite based on tannic acid-metal for near-infrared fluorescence/magnetic resonance bimodal imaging-guided combinational therapy in Alzheimer’s disease

Scheme 1: Nanocomposite preparation and therapeutic mechanism

TrafficCAM: A Versatile Dataset for Traffic Flow Segmentation

Published online in 2024; journal issue: February 2025.

Single-Shot Plug-and-Play Methods for Inverse Problems

Figure 1: Single-Shot Plug-and-Play pipeline

TrafficMOT: A Challenging Dataset for Multi-Object Tracking in Complex Traffic Scenarios

Figure 2: Annotated object classes in TrafficMOT

Anatomical Structure-Guided Medical Vision-Language Pre-training

Revitalizing Multivariate Time Series Forecasting: Learnable Decomposition with Inter-Series Dependencies and Intra-Series Variations Modeling

ConSlide: Asynchronous Hierarchical Interaction Transformer with Breakup-Reorganize Rehearsal for Continual Whole Slide Image Analysis

Figure 2: Continual whole slide image analysis framework

MulGT: Multi-task Graph-Transformer with Task-aware Knowledge Injection and Domain Knowledge-driven Pooling for Whole Slide Image Analysis

A deep learning model for detection of Alzheimer’s disease based on retinal photographs: a retrospective, multicentre case-control study

Multitask Learning-Driven Volume and Slice Level Contrastive Learning for 3D Medical Image Classification

Multi-Granularity Cross-modal Alignment for Generalized Medical Visual Representation Learning

Dual-Teacher++: Exploiting Intra-Domain and Inter-Domain Knowledge With Reliable Transfer for Cardiac Segmentation

DoFE: Domain-Oriented Feature Embedding for Generalizable Fundus Image Segmentation on Unseen Datasets

REFUGE Challenge: A Unified Framework for Evaluating Automated Methods for Glaucoma Assessment from Fundus Photographs

Learning from Extrinsic and Intrinsic Supervisions for Domain Generalization

Dual-Teacher: Integrating Intra-domain and Inter-domain Teachers for Annotation-Efficient Cardiac Segmentation

Towards Cross-Modality Medical Image Segmentation with Online Mutual Knowledge Distillation

RMDL: Recalibrated Multi-instance Deep Learning for Whole Slide Gastric Image Classification

Patch-Based Output Space Adversarial Learning for Joint Optic Disc and Cup Segmentation

Boundary and Entropy-Driven Adversarial Learning for Fundus Image Segmentation

Uncertainty-Aware Self-ensembling Model for Semi-supervised 3D Left Atrium Segmentation

Agent with Warm Start and Active Termination for Plane Localization in 3D Ultrasound

Unsupervised Retina Image Synthesis via Disentangled Representation Learning