Klasifikasi Gerakan Jari Tangan Berdasarkan Sinyal Electromyogram Pada Lengan

Abstract

An electromyogram is a recording of muscle activity. These signals have been used both for medical diagnosis and engineering such as finger motion detection in healthy people and rehabilitation patients. Many studies have been conducted to map the relationship between electromyogram and finger movements, one of which is the relationship between the number of channels used and the complexity of the system. The number of channels used is directly proportional to the complexity of a system. The more complex the system, the heavier the data processing is so that it requires greater resources. Therefore, this study focuses on the construction of a classification system for human finger movements using fewer channels. The number of channels used in this study is 4. Root Mean Square is applied in a sliding window as feature extraction. The classifier used is the artificial neural network. System validation is done with 10-fold cross-validation. The test results of the average accuracy value for the thumb, index finger, middle finger, ring finger, little finger, grip, and relaxation were 89%, 90%, 93%, 95%, 93%, 94%, and 91% respectively which can be said to be quite good considering the number of channels relatively few compared to previous studies.