Gastrointestinal Disease Classification And Analysis Using GI-Net Model

Abstract

Detecting gastrointestinal illnesses accurately is crucial to early cancer detection and treatment. In spite of this, manual analysis is time-consuming, requiring the assistance of a gastrointestinal. A multi-class classification framework for screening gastrointestinal diseases is proposed that is efficient and robust. A neural network known as gastrointestinal GI-Net was developed to extract features that could differentiate between normal and diseased images taken by endoscopy device. In order to achieve the most optimal classification network, a variety of optimization techniques are used. For the classification network to be more effective, the framework can handle the challenges present in the dataset. It is 88% accurate in diagnosing using unseen endoscopy images. In comparison with other deep learning networks, the developed architecture is highly effective. Compared to other networks, in limited computation environments, the proposed network is likely to perform better.