Convolutional Neural Network in Motion Detection for Physiotherapy Exercise Movement

Laistulloh, Dika Fikri and Handayani, Anik Nur and Asmara, Rosa Andrie and Taw, Phillip (2024) Convolutional Neural Network in Motion Detection for Physiotherapy Exercise Movement. Knowledge Engineering and Data Science, 7 (1). pp. 27-39. ISSN 2597-4637

Text 47747-168205-3-PB.pdf Download (1MB)

Abstract

Physiotherapy focuses on movement and optimal utilization of the patient's potential. Exercise Therapy is a physiotherapy procedure that specifically focuses exercises on active and passive movements. Cerebral Palsy (CP) patients are one of the sufferers of motor disorders of the upper extremities. Cerebral Palsy (CP) patients suffer from disorders in motor functions of the upper extremities. Physiotherapy Exercise Movement has 4 categories of movement exercises for the therapy of people with upper extremity body disorders: Elbow flexor strengthening in sitting using free weights, lifting an object up, reaching diagonally in sitting, and reaching from a low surface to a high surface. By taking 4 categories of motion movements in exercise therapy, data were taken using normal child subjects as standard movements, which then became a reference for CP child therapy. The limitations of therapy in physical care prompted researchers to investigate the use of image processing as input to Human Computer Interaction (HCI) in the process of motion detection-based therapy. In research using Deep learning as a classifier, namely using the CNN Model (Inception V3, Resnet152, and VGG16 architectural models). The results obtained by the CNN (Inception V3) model have the best performance with an accuracy percentage of 98%.

Item Type:	Artikel Umum
Subjects:	T Technology > TK Electrical engineering. Electronics Nuclear engineering
Divisi / Prodi:	Faculty of Industrial Technology (Fakultas Teknologi Industri) > S1-Electrical Engineering (S1-Teknik Elektro)
Depositing User:	M.Eng. Alfian Ma'arif
Date Deposited:	21 Apr 2025 01:37
Last Modified:	21 Apr 2025 01:37
URI:	http://eprints.uad.ac.id/id/eprint/83071

Actions (login required)

View Item