Deep Learning analysis using ResNet for Early Detection of Cerebellar Ataxia Disease

Abstract

Cerebellar Ataxia disease (CA) is one of the neurological diseases that makes the critical health issues in affected patients. For this goal, disease prediction should closely study the premotor stage of Cerebellar Ataxia disease. A novel deep-learning algorithm is used to determine whether a person has Cerebellar Ataxia disease based on promoter traits. In addition to recognizing the CA, we also discuss the feature importance of the Boosting-based CA detection process. The research investigated many tests to detect CA, like Rapid Eye Movement and slow activity movements or wrong movements. The proposed research model is based on a collected dataset, including 195 patients with regular and affected persons. The different images are classified using the various movement factors. This research designed the ResNet50 model, which gives an average accuracy of 87.5%.