Enhanced Pelican Optimization Algorithm with Deep Learning-Driven Mitotic Nuclei Classification on Breast...

Area of Science:

• Medical Imaging and Pathology
• Artificial Intelligence in Medicine
• Computational Biology

Background:

• Accurate breast cancer diagnosis is crucial for effective treatment, with histopathological inspection of mitotic nuclei being a key factor.
• Manual detection of mitotic nuclei by pathologists is subjective and time-consuming.
• Deep learning (DL) offers a promising automated alternative for accurate and efficient classification of mitotic nuclei.

Purpose of the Study:

• To develop and validate an Enhanced Pelican Optimization Algorithm with Deep Learning-Driven Mitotic Nuclei Classification (EPOADL-MNC) for breast histopathology images.
• To automate the classification of mitotic and non-mitotic cells in breast tissue samples.
• To improve the accuracy and efficiency of mitotic nuclei detection compared to existing methods.

Main Methods:

• Utilized the ShuffleNet model for feature extraction from histopathology images.
• Employed the Enhanced Pelican Optimization Algorithm (EPOA) for hyperparameter tuning of the ShuffleNet model.
• Integrated an adaptive neuro-fuzzy inference system (ANFIS) for the final classification and detection of mitotic cell nuclei.

Main Results:

• The developed EPOADL-MNC technique demonstrated superior performance in classifying mitotic nuclei.
• Simulations confirmed the improved detection capabilities of the EPOADL-MNC system.
• Achieved a maximum classification accuracy of 97.83%, outperforming existing deep learning techniques.

Conclusions:

• The EPOADL-MNC technique provides an accurate and efficient automated solution for mitotic nuclei classification in breast histopathology.
• This deep learning approach enhances diagnostic accuracy and reduces the burden on pathologists.
• The method holds significant potential for improving breast cancer prognosis and diagnosis.