Upper Limb Assessment in Virtual Reality
People who have suffered a stroke may face upper limb motor impairments. Many daily life’s tasks involve the use of the upper limbs. Therefore, rehabilitation is necessary to improve motor ability and recovery of stroke patients. However, stroke patient may become bored and unmotivated when doing th...
Saved in:
| Main Author: | |
|---|---|
| Format: | Thesis |
| Published: |
2021
|
| Subjects: | |
| Tags: |
Add Tag
No Tags, Be the first to tag this record!
|
| Summary: | People who have suffered a stroke may face upper limb motor impairments. Many daily life’s tasks involve the use of the upper limbs. Therefore, rehabilitation is necessary to improve motor ability and recovery of stroke patients. However, stroke patient may become bored and unmotivated when doing the same rehabilitation tasks repeatedly for a long duration. Most rehabilitation exercises used devices such as Microsoft Kinect and leap sensor which were not equipped with immersive technology. The rehabilation can be more labour intensive and time consuming for the physiotherapist to monitor the patient during rehabilitation without the use of biosignal technology. Moreover, biosginals required a pre-processing method before a signal value can be obtained. The whole process could be time consuming. This study presents an evaluation of the performance based on a virtual reality (VR) upper limb assessment, feedback from forty human subjects and biosignal image classification by using deep learning. Therefore, an assessment was created from Unreal Engine 4 (UE4), the Virtual Reality (VR) Upper Limb Assessment. There were four VR games namely ‘Pick and Place’ (PNP) game, ‘Mirror Pick and Place’ (MPNP) game, ‘Hit the Ball’ (HTB) game, and ‘Wall Climbing’ (WC) game. Before performing the assessment, the precision and accuracy of the virtual reality devices need to be quantified. Pre-assessment was performed to show the upper limb length measurements and the position and rotation of the upper limb were estimated correctly. The upper limb assessment evaluates the performance of the user and gives feedback to the user. During the assessment, the users were attached with the biosignal devices namely, electroencephalogram (EEG), electromyographic (EMG), and electrocardiographic (ECG). The three raw biosignals were collectively represented as images and were used to train deep learning models (namely, convolutional neural network and long-short term memory). Based on the four games, the stroke group archived a better score in the MPNP game compared to the other games. A NASALoad Task Index shows that HTB game has the highest task load while MPNP game has the lowest task load. Feedback from users shows that the system was easy to use, full of fun and helpful. The classification performance of the deep neural network in classifying the biosignals is highly accurate and precise. |
|---|