Background: This research investigates the application of machine learning and deep learning techniques for classifying gastrointestinal (GI) disorders using the Kvasir dataset, which contains annotated endoscopic images across eight diagnostic categories.The study begins with traditional approaches such as logistic regression and support vector machines and then progresses to more advanced deep learning models, including Mobile-NetV3, VGG19, and Inception model version 3 (InceptionV3), to evaluate their effectiveness in medical image analysis.Methods: Performance metrics such as accuracy, loss trajectories, and confusion matrices are employed for model assessment.The results indicate that InceptionV3 demonstrated the strongest generalization capability and the most consistent performance across all disease diagnostic classes.While classical models provide a foundational framework, they encounter limitations in capturing the complex visual features inherent in medical images.Results: In contrast, deep learning models, particularly InceptionV3 and VGG19, exhibit superior accuracy and reliability in diagnosing GI conditions.Conclusion: Overall, the findings underscore the promising role of advanced deep learning architectures in enhancing diagnostic precision for complex medical imaging tasks.
Kumar et al. (Tue,) studied this question.