Hybrid Characteristics are used in an Efficient Framework for the Identification of Gastrointestinal Diseases

Medical fields frequently use machine learning (ML) algorithms for disease classification and diagnosis. These algorithms base their judgments on features that have been retrieved from images. To achieve high classification accuracy choosing proper feature extraction methods and Machine Learning (ML) algorithms is crucial. This study discovers a dearth of knowledge on computer learning techniques for GI disease classification through the literature review. To address this problem, five classes of GI images are gathered from a freely accessible mdi dataset and extracted textural features like RGB and HSV (Hue, Saturation, and Value) color characteristics GLCM Matrix shape data from HOG. These characteristics are integrated, and the size was decreased by Principal Components Analysis (PCA). Last but not least, the chosen features are classified using the well-known ML classifier OAA-SVM -One Against All-Support Vector Machine, achieved classification precision, recall, FI-Score values, and accuracy of 98.8%, 95.44%, and 98.64% respectively.