Gastrointestinal Disease Diagnosis through Deep Features Extracted from Endoscopic Imaging

This study explores the efficiency of deep feature extraction combined with machine learning classifiers for precise gastrointestinal disorder diagnosis. Features were extracted from gastrointestinal images using DenseNet-201, and the performance was compared across several machine learning algorithms. The kernel support vector machine (SVM - K) classifier outperformed other methods, demonstrating superior accuracy in disease diagnosis. Additionally, we evaluated alternative pre-trained convolutional neural network (CNN) models, further enhancing diagnostic accuracy. These findings contribute to improving gastrointestinal disease diagnosis and patient care.