Generous Approach for Diagnosis and Detection of Gastrointestinal Tract Disease with Application of Deep Neural Network
Gunjan Sharma, Vatsala Anand, Sheifali Gupta
Abstract
An essential component of the human digestive system is the colon, sometimes addressed as the big intestine at times. It is a lengthy, tube-like organ that follows the small intestine at the end of the digestive system. Polyps are noncancerous growths that develop on the large intestine's inner lining in the colon. These polyps can further result in cancer. The management of colon diseases and the improvement of outcomes depend critically on early identification and proper medical care. For early detection of colon disease, CNN plays an important role as with this the colon disease can be further detected. This work proposes a CNN model based on the EfficientNetB5 pre-trained model to classify the colon disease images into four classes. A Wireless Capsule Endoscopy (WCE) curated colon disease dataset has been employed for training and testing the model. The model is evaluated with a training accuracy of 99.01% and a validation accuracy of 97.85%. The model has shown very satisfactory results for Precision and Recall for the classification of colon disease. This research can be further employed in the medical area for the early detection of Colorectal Cancer.