Computer-aided polyp segmentation is an important activity that assists gastroenterologists in evaluating and removing abnormal tissue from gastrointestinal system. Disease polyps arise mostly in colorectal area of gastrointestinal system and in mucous membrane, that includes micro-abnormal tissue protrusions that raise risk of incurable illnesses such as cancer. Hence, early evaluation of polyps can reduce likelihood of polyps developing into cancer, such as adenomas, that can develop into cancer. Deep learning-based diagnostic tools are critical for early illness detection. To segregate polyps using colonoscopy frames, a deep learning approach known as Mask R-CNN is presented. Mask R-CNN having RPN-CNN backbone is a modified version of standard F-CNN having 3 stages: feature extraction, ROI pooling, and segmentation for severity analysis. Mask R-CNN technique outperforms previous deep learning techniques in segmentation. trials utilized a single dataset that is obtained from big intestine of gastrointestinal system via a colonoscopy technique. predicted technique exceeds having a mean Dice of 98 percentage, accuracy of 96.6 percentage, precision of 97 percentage, recall of 96.6 percentage, and an F1-score of 92 percentage.

Segmentation, Colorectal Polyps, Feature Extraction, Neural Network, Pooling Process