DCNN with Barnacles Mating Optimizer for Performance Analysis of Classification for Liver Disease

Currently in development is a technique for liver disease prediction meant to help in early identification of human liver diseases. Found in the right upper quadrant of the belly, the liver is a critical organ Effective treatment of liver illness depends on a preliminary diagnosis of it. Early illness identification using conventional approaches still presents difficulties in medical research, nevertheless. Methodologies of machine learning can help to find latent diagnostic information and provide precise conclusions. Two different approaches of algorithms to solve the challenging problem of liver disease categorisation are presented in this work. Based on conventional machine learning, the first method sorts objects using texture data. We do this using traditional machine learning techniques like automated texture analysis and supervised learning. Using supervised learning methods helps us to produce suitable picture classes reflecting the same situation, therefore improving our capacity to identify liver tissue flaws. Using machine learning, our suggested technique quantifies asymmetric patterns in CT images quite successfully. To rapidly extract characteristics, we used the Barnacles Mating Optimizer (BMO) method. To generate its findings, our approach combines a deep convolutional neural network (DCNN) with semantic image segmentation (SIS). The model employs a novel guided attention strategy to encode high-density maps. The trained model classifies CT images into five different disease categories. The compelling results confirm the model's effectiveness. The findings suggest that texture analysis methods can be utilized to detect and classify abnormalities in the human liver, assisting medical physicists and radiologists in predicting the severity and progression of liver diseases. According to experimental data, the proposed DCNN-BMO model achieves a high accuracy of 98.04%.