Optimised feature selection-driven convolutional neural network using gray level co-occurrence matrix for detection...

Area of Science:

• Oncology
• Medical Imaging
• Machine Learning

Background:

• Cervical cancer remains a significant global health threat, disproportionately affecting women in developing regions.
• While incidence has decreased in industrialized nations, progress is slower in developing countries.
• Early detection through screening is crucial for managing cervical cancer, which is often diagnosed late due to its slow progression.

Purpose of the Study:

• To evaluate the effectiveness of advanced image analysis techniques for cervical cancer detection.
• To compare the performance of different machine learning classifiers in identifying early-stage cervical cancer.
• To enhance the precision of cervical cancer screening methods.

Main Methods:

• Feature selection was performed using the Gray Level Co-occurrence Matrix (GLCM) technique.
• Classification models, including Convolutional Neural Network (CNN), Support Vector Machine (SVM), and Autoencoder, were employed.
• The GLCM technique was integrated with these classifiers for comparative analysis.

Main Results:

• The GLCM-CNN classifier achieved the highest precision among the tested methods.
• This indicates superior performance in accurately identifying cervical cancer features.
• The study highlights the potential of GLCM-CNN for improved diagnostic accuracy.

Conclusions:

• The GLCM-CNN classifier shows significant promise for enhancing cervical cancer screening accuracy.
• Integrating GLCM with CNN offers a powerful tool for early detection and diagnosis.
• Further research could lead to more effective cervical cancer management strategies, particularly in underserved regions.