Solving the Small Sample Problem in Pathological Image Segmentation: Approaches for Accurate Tissue and Nucleus Analysis
数学专题报告
报告题目(Title)：Solving the Small Sample Problem in Pathological Image Segmentation: Approaches for Accurate Tissue and Nucleus Analysis
报告人(Speaker)：贵鹿颖（南京理工大学）
地点(Place)：后主楼1124
时间(Time)：2025年11月9日（周日）15:00-16:00
邀请人(Inviter)：段玉萍
报告摘要
Pathological image segmentation faces substantial difficulties due to the variability in tissue structures across different diseases and the challenge of obtaining ground-truth annotations, which often transforms it into a small-sample problem. Large models tailored for natural images typically struggle with pathological data, while those specific to pathology rarely tackle segmentation tasks. In response, we introduce methods to overcome these issues, enabling precise segmentation of tissues and nuclei in pathological images. This facilitates more interpretable and advanced analysis of histopathological data.
主讲人简介
贵鹿颖，南京理工大学数学与统计学院副教授，主要从事图像处理的模型与算法研究，包括基于数据和模型共同驱动的图像分割算法，病理图像分析等。主持国家自然科学基金青年项目一项、面上项目一项，参与科技部重点专项及国家自然科学基金重点项目等。
