Cervical cancer histology image identification method based on texture and lesion area features

Area of Science:

• Medical Imaging
• Computational Pathology
• Oncology

Background:

• Cervical cancer detection relies on accurate histological analysis.
• Automated methods can potentially enhance diagnostic accuracy and efficiency.

Purpose of the Study:

• To develop and validate an automated approach for cervical cancer detection from histology images.
• To improve the accuracy of cervical cancer recognition through image processing and machine learning.

Main Methods:

• Image preprocessing to reduce noise and background interference.
• Texture feature extraction using Grey Level Co-occurrence Matrix (GLCM).
• Image segmentation via K-means clustering and Marker-controlled watershed algorithm, followed by Support Vector Machine (SVM) classification.

Main Results:

• The automated approach demonstrated effectiveness in recognizing cervical cancer.
• Experimental validation confirmed the feasibility of the proposed method.
• Texture features and lesion area calculations contributed to accurate tissue investigation.

Conclusions:

• The developed automated system is effective and feasible for cervical cancer detection.
• The integration of GLCM, segmentation algorithms, and SVM shows promise for automated histological analysis.
• This approach offers a potential improvement in the accuracy of cervical cancer recognition.