Related Experiment Video
Updated: Mar 27, 2026
06:23
Detection of Nuclear Blebbing and DNA Leakage in Mammalian Cells by Immunofluorescence
Published on: January 17, 2025
1.3K
Unsupervised HEp-2 mitosis recognition in indirect immunofluorescence imaging.
Summary
This study introduces an unsupervised method for recognizing mitotic cells in HEp-2 images, crucial for diagnosing autoimmune disorders. The approach effectively handles class imbalance, improving computer-aided diagnosis accuracy.
Area of Science:
- Medical Imaging
- Computational Biology
- Immunology
Background:
- Automated recognition of HEp-2 mitotic cells in immunofluorescence (IIF) images is vital for computer-aided diagnosis of autoimmune disorders.
- Accurate identification of mitotic cells aids in assessing sample quality and diagnosing challenging cases.
Purpose of the Study:
- To develop a completely unsupervised approach for HEp-2 mitotic cell recognition.
- To address the challenge of mitotic/non-mitotic class imbalance in automated cell recognition.
Main Methods:
- The proposed technique involves an unsupervised method for HEp-2 mitotic cell recognition.
- It automatically selects candidate cells and uses clustering based on texture to identify mitotic cells.
- A subsequent clustering stage differentiates between positive and negative mitoses.
Main Results:
- The unsupervised approach effectively overcomes the class imbalance problem inherent in identifying rare mitotic cells.
- Experimental results on public IIF images demonstrate competitive performance compared to existing methods.
Conclusions:
- The developed unsupervised method offers a promising solution for automated HEp-2 mitotic cell recognition.
- This technique can enhance the accuracy and efficiency of computer-aided diagnosis for autoimmune diseases.
