A Review of Computational Methods for Cervical Cells Segmentation and Abnormality Classification

Area of Science:

• Oncology
• Medical Imaging
• Computer Science

Background:

• Cervical cancer is a leading global cancer in women, with current screening methods facing challenges like lengthy workflows and ambiguity.
• These challenges are particularly pronounced in low-income regions, contributing to higher mortality rates.
• There is a growing need for efficient and accurate cervical cancer screening solutions.

Purpose of the Study:

• To provide an overview of cervical cancer and its screening procedures.
• To analyze computational methods for automated cervical cell analysis, focusing on quality assessment, segmentation, and classification.
• To support the development of next-generation computer-aided diagnosis (CADx) systems for cervical screening.

Main Methods:

• Extensive literature review of computational methods for cervical cell analysis.
• Focus on automated quality assessment, segmentation, and classification techniques.
• Critical discussion of existing methods and their relevance to practical application.

Main Results:

• Identification and analysis of key computational approaches for cervical cell analysis.
• Evaluation of methods for automated quality assessment, segmentation, and classification.
• Discussion of the strengths and limitations of current computer-aided methods.

Conclusions:

• Computer-aided diagnosis systems offer significant potential to enhance cervical cancer screening efficiency and accuracy.
• Further research into automated tools is crucial for overcoming current screening limitations, especially in resource-limited settings.
• The review highlights future research directions for developing advanced CADx systems in cervical cancer screening.