[Polar coordinates representation based leukocyte segmentation of microscopic cell images]

Area of Science:

• Biomedical Image Analysis
• Computational Biology
• Cellular Microscopy

Context:

• Accurate segmentation of leukocytes is crucial for hematological analysis and disease diagnosis.
• Overlapping cells in microscopic images present a significant challenge for automated analysis.
• Existing segmentation methods often struggle with complex overlapping cell structures.

Purpose:

• To develop an automated algorithm for precise segmentation of overlapped leukocytes in microscopic cell images.
• To improve the accuracy and robustness of cell segmentation in challenging imaging conditions.
• To provide a reliable tool for quantitative analysis of leukocyte populations.

Summary:

• The proposed algorithm utilizes histogram smoothing of the saturation channel to identify the nucleus, enabling initial segmentation.
• A circular region is defined based on the nucleus's radius, and leukocyte edges are represented using polar coordinates.
• Overlapping regions are detected by analyzing changes in polar angles, and edges are reconstructed using gradient information for accurate leukocyte extraction.

Impact:

• The algorithm demonstrates high performance in recall ratio, precision ratio, and pixel error ratio for overlapped leukocyte segmentation.
• This method can enhance the efficiency and accuracy of automated cell counting and analysis in clinical and research settings.
• Improved leukocyte segmentation facilitates more reliable diagnostic insights from microscopic cell imaging.