Detection and Classification of Colorectal Polyp Using Deep Learning.

Area of Science:

• Medical Imaging
• Artificial Intelligence in Medicine
• Oncology

Background:

• Colorectal cancer (CRC) is a leading global health concern, necessitating early and precise diagnosis.
• Colonoscopy is crucial for detecting colorectal polyps, but manual analysis is error-prone.
• Existing automated methods struggle with overfitting and gradient vanishing.

Purpose of the Study:

• To develop an automated deep learning model for accurate detection and classification of colorectal polyps.
• To address limitations of existing models, such as overfitting and gradient vanishing.
• To enhance the diagnostic accuracy of colonoscopy through image processing and deep learning.

Main Methods:

• Utilized guided image filtering and dynamic histogram equalization for colonoscopy image enhancement.
• Employed a Single Shot MultiBox Detector (SSD) for efficient polyp detection and classification.
• Incorporated fully connected layers with dropouts in a convolutional neural network (CNN) architecture.

Main Results:

• The proposed CNN-based model demonstrated superior performance compared to existing methods on benchmark datasets.
• Achieved a high accuracy of 92% in detecting and classifying colorectal polyps from colonoscopy images.
• Successfully mitigated overfitting and gradient vanishing issues common in deep learning models.

Conclusions:

• The developed deep learning model offers a promising solution for automated and accurate colorectal polyp diagnosis.
• Early and precise polyp detection can significantly reduce the impact of colorectal cancer.
• This approach has the potential to improve patient outcomes by enhancing colonoscopy effectiveness.