Differential evolution based advised SVM for histopathalogical image analysis for skin cancer detection.

Area of Science:

• Digital Pathology
• Computational Biology
• Medical Image Analysis

Background:

• Automated detection of cancerous tissue in histopathological images presents a significant challenge in cancer diagnosis.
• Accurate identification of malignant cells is crucial for effective treatment planning and patient outcomes.

Purpose of the Study:

• To propose a novel pattern recognition method for the automated identification and classification of cancerous tissues in histopathological images.
• To enhance the accuracy and efficiency of cancer detection in digital pathology.

Main Methods:

• Feature extraction was performed using a combination of wavelet features, intensity-based statistical features, and autoregressive parameters.
• Differential evolution was employed for feature selection to reduce dimensionality.
• A self-advised support vector machine (SVM) was utilized for feature evaluation and image classification.

Main Results:

• The proposed system demonstrated promising comparative results on a dataset of 150 histopathological images.
• An average diagnostic accuracy of 89.1% was achieved, indicating the effectiveness of the developed method.
• The combination of advanced feature extraction and selection techniques contributed to improved classification performance.

Conclusions:

• The developed pattern recognition method shows significant potential for automated cancer detection in histopathology.
• The proposed approach offers a robust and accurate solution for analyzing histopathological images, aiding in cancer diagnosis.
• Further research can explore larger datasets and diverse cancer types to validate and refine the system.