Detection of Colorectal Polyps from Colonoscopy Using Machine Learning: A Survey on Modern Techniques

Area of Science:

• Medical Imaging
• Artificial Intelligence
• Oncology

Background:

• Colorectal polyps are precursors to colorectal cancer (CRC), a leading cause of cancer mortality.
• Physician fatigue and experience gaps can lead to missed polyp diagnoses during colonoscopies.
• Early and accurate polyp detection is crucial for preventing CRC progression.

Purpose of the Study:

• To survey recent advancements in deep learning for colorectal polyp detection and classification.
• To analyze common challenges and benchmark datasets in polyp detection research.
• To identify trends and gaps for future research in AI-assisted colonoscopy.

Main Methods:

• Comprehensive literature review of AI-based polyp detection methods.
• Analysis of benchmark datasets and evaluation metrics used in recent studies.
• Categorization of common challenges including data scarcity and image artifacts.

Main Results:

• Deep learning shows promise for improving polyp detection accuracy.
• Key challenges remain, including insufficient training data and issues with white light reflection and blur.
• Current methods vary in their approaches to building polyp detectors.

Conclusions:

• Despite progress, significant challenges hinder the widespread clinical adoption of AI for polyp detection.
• Further research is needed to address data limitations and improve robustness against image artifacts.
• Identifying trends and gaps is essential for guiding future development in AI-assisted colonoscopy.