Development of Attention Guided Hybrid Network with Adaptive FCM-aided Segmentation for Predicting the Liver Disease

Liver Disease (LD) is considered as one of the primary death causes around the world affecting a vast amount of people. A group of factors trouble the liver leads to make dysfunction. The detection of this condition becomes tiresome and costly. There are numerous disorders related to the human liver some of them are complex to recognize by employing the data transmitted among the doctor and patient. Inspired by the better knowledge of Artificial Intelligence (AI) in machines, this work concentrated to discover the strategy that can forecast the existence of LD in specific subjects with improved accuracy on the basis of distinct input attributes. Accurate and fast LD prediction permits timely and efficient treatments. Due to the significance of LD and the increase in the amount of people who are troubled by this disease; the work concentrated the LD via deep learning method. Hence, this work proposes an automated model of LD prediction using the adaptive technique. Initially, the benchmark datasets are considered for gathering the required set of images used for subsequent processes. Further, the gathered input images are subjected to the segmentation process, which is accomplished by Optimized Fuzzy C-Means (OFCM). In this OFCM, the parameters are optimally chosen by using the Amplified Hippopotamus Optimization (AHO) algorithm. Finally, the segmented image is fed into the model of Attention Guided Hybrid Network (AGHNet), in which the Graph Convolutional Neural Network (GCNN) is integrated with the Recurrent Neural Network (RNN) for predicting the LD. Further, the proposed work validates with multiple metrics to deliver the desired outcome. Compared with traditional approaches, the recommended prediction model outperforms with the superior results of early time disease diagnostic approach.