Improving the diagnosis of Gastrointestinal Diseases using a modified AlexNet

Abstract

The digestive system is a long tube that runs from the mouth to the anus and is made up of various hollow organs. Patients with gastrointestinal diseases must have appropriate treatment and an early diagnosis to recover. Biomedical applications have recently witnessed a surge in research activity in detecting gastrointestinal tract illnesses by endoscopic image classification. However, previous research suggested that additional work is necessary due to the difficulty of distinguishing certain gastrointestinal disease classifications from others. Deep learning algorithms can diagnose gastrointestinal diseases like famous diseases, such as ulcerative colitis, polyps, and esophagitis. Convolutional neural networks are computer-aided diagnosis techniques' most famous deep learning branches. One of the most often used convolutional neural network models for classification techniques is AlexNet. An enhanced AlexNet technique to improve the multi-class diagnosis of gastrointestinal disorders is presented in this suggested work. In this approach, the proposed AlexNet, proposed as T-AlexNet, proposes employing the Tanh activation function instead of the Relu activation for more complicated multi-class disease datasets. T-AlexNet has better performance evaluation parameters in accuracy, sensitivity, and specificity than normal AlexNet, AlexNet enhancement techniques, and other recent approaches for gastrointestinal disease multi-class diagnosis. From experimental results. The accuracy obtained is 0.995, better than standard AlexNet, which obtained an accuracy of 0.94, so accuracy was enhanced by 5%.