M.S. Student at Fuzhou University · Applying for Fall 2027 PhD programs
My research focuses on reliable and label-efficient learning for medical image analysis, with computational pathology as the primary application domain. I study how to extract fine-grained visual evidence, learn from weak supervision and multimodal information, and improve model generalization across devices, institutions, and clinical settings.
How can we extract clinically meaningful fine-grained evidence from complex medical images?
I investigate cell-level detection and representation learning in computational pathology, including cervical cytology and whole-slide image analysis.
How can we learn effectively when dense medical annotations are scarce?
I use textual descriptions and other inexpensive supervision signals to complement or replace costly dense annotations for medical image analysis.
How can medical vision models remain reliable across devices, institutions, and clinical settings?
I study domain generalization and robust representations under scanner, center, preprocessing, and other distribution shifts.
IEEE International Conference on Multimedia & Expo · First Author
TG-WSPIS uses three granularities of text descriptions and soft-alignment consensus to model heterogeneous cellular distributions, achieving near-fully-supervised pathology segmentation with text-only weak supervision.
First Author · Manuscript under Review
We study single-source domain generalization for medical object detection, focusing on instance-level response heterogeneity under distribution shifts. The method improves robustness across scanners and datasets without additional inference-time computation.
Neurocomputing · First Author
CytoSet refines and spatially groups detection queries while coupling ordered center alignment with prediction quality, improving Bethesda-grade consistency and reducing high-risk cell downgrading errors in fine-grained detection.
Chinese Conference on Pattern Recognition and Computer Vision · Fourth Author
CRAFT estimates sample-level text reliability through cross-modal consistency and uses a counterfactual branch to reduce unreliable language influence, mitigating language-dominated prediction bias and improving multimodal fusion robustness.
Chinese Conference on Pattern Recognition and Computer Vision · Fourth Author
DT-VG combines hierarchical text encoding with physics-aware structural constraints to capture multi-segment semantic dependencies and motion consistency, predicting continuous and geometrically precise trajectories from a single static drone image.
IEEE Transactions on Multimedia · Sixth Author
Image-grounded morphology steering uses frozen text prototypes for semantic compatibility and bounded residuals at the final decoder layer to strengthen diagnostic morphology evidence, improving fine-grained discrimination and cross-domain localization in deformable detectors.
I aim to develop reliable and label-efficient learning methods for medical image analysis, with a particular focus on computational pathology under weak supervision and distribution shifts.
My current research centers on three connected questions: how to learn fine-grained visual evidence from limited annotations or multimodal supervision; how to maintain robustness across scanners, institutions, and evolving data distributions; and how to support medical predictions with evidence that can be localized, quantified, and verified.
During my PhD, I hope to further investigate trustworthy medical vision in weakly annotated and multi-center settings, particularly by connecting visual evidence in pathology with medical text, modeling cross-domain variation, and developing risk-aware learning strategies. My goal is to improve the auditability and cross-center applicability of medical imaging models and establish a stronger foundation for downstream clinical validation.
Cervical cytology slides lack the relatively stable spatial structure and contextual information found in histopathology, while abnormal cells are sparse, morphological differences are subtle, and whole-slide images are extremely large. The core challenge in automated TCT screening is to identify diagnostically meaningful abnormal cells within vast numbers of normal cells and complex backgrounds, then form a reliable slide-level screening result.
I lead the visual screening pipeline from TCT WSI processing and cell detection to abnormal-cell evidence aggregation, slide-level prediction, and evaluation, including experimental dataset construction, model training, result analysis, and screening-threshold selection.
In internal validation, the model achieved approximately 50% specificity at a high-recall operating point of 95% sensitivity, enabling effective triage of many negative slides while prioritizing the detection of abnormal cases. In a screening setting with an abnormal-case prevalence of about 10%, this corresponds to an estimated 45% of all slides entering a low-risk queue, shifting manual review from exhaustive screening toward prioritized review of high-risk cases and key cells.
The project establishes a complete analytical pathway from ultra-high-resolution WSI to cell-level evidence and then to slide-level screening results. This workflow may reduce repetitive review of negative slides in large-scale cervical cytology screening and direct limited human review capacity toward high-risk cases and diagnostically important regions.
I am a master’s student in Computer Technology at Fuzhou University, advised by Fuhai Chen. My research focuses on computational pathology and medical image analysis, particularly cervical cytology, weakly supervised and multimodal learning, and robust generalization across imaging domains.
School of Computer and Data Science · GPA: 3.4 / 4.0
Ranked 2nd in comprehensive evaluation in the major.
Experienced in whole-slide image processing, patch construction, and pathology data organization, supporting both cell-level analysis and WSI-level research workflows.
Able to independently implement, train, evaluate, and analyze medical vision models, with an emphasis on stable and reproducible experimentation.
Experienced in model inference optimization and engineering deployment, while maintaining shared laboratory GPU servers and coordinating computing resources for research.
I am currently seeking PhD opportunities for Fall 2027. I would be glad to connect with researchers working on computational pathology, medical image segmentation and object detection, weakly supervised learning, multimodal learning, and domain generalization, and I am open to potential research collaborations and PhD opportunities.