Abstract

Gait prediction plays a vital role in determining normal and abnormal walk patterns.  Walking pattern is not given much importance by many persons but it is our basic transportation process. A person who faces the inability to walk can bring severe transformation to his/her life. They must always depend upon other people for their day to day activities and also have serious health issues. In some cases, a person with an abnormal gait pattern can walk normally without experiencing any of the symptoms and it can be identified only when the person meets with an injury or pain and it can lead to health issues such as cardiovascular problems, mental illness problem and musculoskeletal problems.Thus, gait prediction must be carried out to identify the abnormal walk pattern, providing proper treatment and lead a healthy life. The sportsperson should also carry out gait prediction methods to help them run effectively. This prediction helps in the diagnosis of the cause for ache, estimation, and classificationof observed abnormalities leading to further treatment.  A method to determine and classify the normal and abnormal gait state based on the Convolution neural network classifier algorithm is proposed. Further usingCNN and SVM the gait abnormalities can be classified as Freezing of gait, Brady Kinesia, Tremor, Ataxic gait, myopathic gait, and muscle atrophy. A comparison for accuracy was performed betweenalgorithms from results obtained from two algorithms. Since the prediction is highly accurate, accurate diagnosis andright treatment by therapists is ensured.