A novel overlapped nuclei splitting algorithm for histopathological images.

Area of Science:

• Digital pathology
• Computational imaging
• Histopathology

Background:

• Nuclei segmentation is crucial for quantitative analysis of histopathological images.
• Overlapping nuclei are a common challenge due to image acquisition and tissue properties.
• Existing 2D methods struggle to represent 3D tissue characteristics, leading to segmentation inaccuracies.

Purpose of the Study:

• To develop and present a novel algorithm for splitting overlapped nuclei in histopathological images.
• To enhance the accuracy of quantitative analysis in digital pathology.

Main Methods:

• Initial segmentation using K-Means algorithm followed by conversion to a binary image.
• Detection of overlapping nuclei using a threshold area value.
• A novel splitting algorithm involving circle drawing and pixel removal to isolate individual nuclei.

Main Results:

• The algorithm was tested on kidney tissue histopathological images.
• Performance was compared against expert segmentation and three related studies.
• The proposed splitting algorithm achieved an accuracy of 84% in segmenting overlapping nuclei.

Conclusions:

• A novel algorithm effectively splits overlapped nuclei in histopathological images.
• Improved cell counting accuracy for histopathological analysis.
• Potential applicability to segmenting other overlapped circular patterns in various fields.