Related Experiment Video
Updated: Mar 7, 2026
Author Spotlight: Enhanced Multiplex Immunofluorescent Microscopy Protocol for Neuroscience Research
Published on: June 21, 2024
Graph-based segmentation of abnormal nuclei in cervical cytology
Ling Zhang1, Hui Kong2, Shaoxiong Liu3
1Department of Biomedical Engineering, School of Medicine, Shenzhen University, Guangdong Key Laboratory for Biomedical Measurements and Ultrasound Imaging, Shenzhen 518060, China; Iowa Institute for Biomedical Imaging, The University of Iowa, Iowa City, IA 52242, USA.
This study introduces a novel graph-search method to accurately segment abnormal cell nuclei in cervical cytology images. The approach enhances diagnostic accuracy by improving the detection of cancerous cells.
Area of Science:
- Medical Imaging
- Computational Pathology
- Cytology
Background:
- Accurate segmentation of cell nuclei is crucial for automated cervical cancer screening.
- Existing methods struggle with segmenting abnormal nuclei, impacting diagnostic sensitivity.
Purpose of the Study:
- To develop a robust method for improving the segmentation of abnormal cell nuclei in cervical cytology.
- To enhance the sensitivity and accuracy of automated analysis of cervical cells.
Main Methods:
- A graph-search algorithm is employed, mapping nucleus contours to a polar coordinate system.
- A cost function integrates nucleus border and region properties, refined by nucleus-cytoplasm spatial constraints.
- Dynamic programming with an iterative approach finds the optimal closed contour.
Main Results:
- The proposed graph-search method demonstrated superior performance in segmenting abnormal nuclei.
- Validation was conducted on the Herlev and H&E stained manual liquid-based cytology (HEMLBC) datasets.
- The method outperformed five other state-of-the-art segmentation approaches.
Conclusions:
- The developed graph-search technique offers a significant improvement for abnormal nucleus segmentation in cervical cytology.
- This method holds promise for enhancing the reliability of automated cervical cancer detection systems.
