An Efficient Blood-Cell Segmentation for the Detection of Hematological Disorders

Area of Science:

• Medical Imaging
• Computational Biology
• Hematology

Background:

• Accurate blood cell segmentation is vital for diagnosing hematological disorders.
• Existing methods face challenges with noise, seed-point detection, and oversegmentation.

Purpose of the Study:

• To propose a novel hybrid ellipse fitting (EF) approach for robust blood cell segmentation.
• To enhance the accuracy and efficiency of detecting hematological disorders.

Main Methods:

• Utilized a Laplacian-of-Gaussian (LoG) based modified highboosting operation for noise reduction.
• Implemented bounded opening followed by fast radial symmetry (BOFRS) for accurate seed-point detection.
• Developed a hybrid least-squares (LS)-based geometric and algebraic EF technique for improved segmentation.

Main Results:

• Achieved more accurate seed-point detection using BO-FRS.
• Demonstrated superior segmentation performance compared to state-of-the-art methods.
• Outperformed existing EF techniques in Dice similarity, Jaccard score, precision, and F1 score.

Conclusions:

• The proposed hybrid EF method offers a computationally efficient and accurate solution for blood cell segmentation.
• This technique shows potential for improved diagnosis and management of hematological disorders.
• The method may also be applicable to other medical and cybernetics applications.