Detection of Colorectal Polyps in Colonoscopy Videos


  • To use deep learning for providing automated
  • solutions for disease diagnosis due to its self-learning capability.
  • To propose a deep learning method for the detection
  • of colorectal polyps from colonoscopy videos using CNN to enhance the accuracy and sensitivity of the existing methods.


  • The proposed work is inspired from the AlexNet Model of CNN and called as containing 6 layers in the first approach, a layer removed in the second approach and then with a layer added.
  • All three were added from 50 epochs having batch size of 64, drop-out rate of 0.5, number of classes-2 and learning rate of 1e-4.


 The automated system can save the traditional manual and time consuming processes.


The proposed CAD system can be used for large scale screening of colorectal polyps at the medical

© Design Innovation All Rights Reserved