BINARY CLASSIFICATION OF ULCERATIVE COLITIS IMAGES USING SUPPORT VECTOR MACHINES (SVMS) WITH THE SQUARE SUM OF SLACK VARIABLES AND SOFTMAX ACTIVATION FUNCTION IN CONVOLUTIONAL NEURAL NETWORKS
Research Scholar, Department of CSE, Annamalai University, Tamil Nadu, India
Assistant Professor, Department of CSE, Annamalai University, Tamil Nadu, India
Professor, Gudlevalleru Engineering College, Andhra Pradesh, India
Ulcerative Colitis (UC) is the prevalence of inflammatory bowel disease in India. It is a chronic disease characterized by periods of time with active inflammation and ulcers in gastrointestinal tract. Several patients with Ulcerative Colitis experience long stretches of inflammation interspersed with flares—periods of active inflammation. The likelihood of developing ulcerative colitis can be influenced by a variety of factors, but the majority can be divided into three categories: biological disposition, environmental exposures, and dysregulated immune reaction. Gastroenterologists use endoscopy and colonoscopy procedures to examine the upper digestive system and inner most lining of the colon visually. However, diagnosis of UC is a difficult task because of its varying traits and variety of patterns.
The goal of this research is to identify Ulcerative Colitis remissions using computational algorithms. For image classification problems, finding the best classifier is more competitive based on high-level deep features of images. In this regard a novel Convolutional Neural Network (CNN) is proposed by using Conv2D and Separable Conv2D classes with SVM Classifier and Softmax activation function for better classification of UC images. Automatic identification and effective learning of prominent features from the UC image is the main intent of utilizing CNN in this research. The proposed approach is compared with existing techniques, the outcome and analysis depict that the research is highly effective.
Keywords: bowel disease, remissions, UC images, CNN Model, SVM Classifier, Conv2D, Separable Conv2D, Softmax activation function.