TY - JOUR
T1 - Selecting Control Menu on Electric Wheelchair Using Eyeball Movement for Diffable Person
AU - Utaminingrum, Fitri
AU - Somawirata, I. Komang
AU - Pengestu, Gusti
AU - Thaipisutikul, Tipajin
AU - Shih, Timothy K.
N1 - Publisher Copyright:
© 2023, Politeknik Negeri Padang. All rights reserved.
PY - 2023
Y1 - 2023
N2 - Each country's number of people with disabilities and strokes increases yearly. Hand defects and stroke make them have limitations in doing activities. It caused their hand has paralyzed. Hence, they find it difficult to do daily activities, such as running a wheelchair, choosing a menu on the screen display, and so on. One solution offered is utilizing eye movement as a navigation tool that can replace the role of the user's hand, so they can run a wheelchair independently or choose a menu selection on display by themselves through the movement of their eyes. Detection of eyeball movements in this study only utilizes a camera as a sensor mounted in front of the user. So that it is more practical and easier to use than if we have to pair an electrooculography sensor in the area around the user's eyes. This research proposed a new approach to detect the five gazes (upward, downward, leftward, rightward, and forward) of the eyeball movements by using Backpropagation Neural Network (BPNN) and Dynamic Line Sector Coordinate (DLSC). Line Sector Coordinate is used to detect the eyeball movement based on the pupil coordinate position. The eyeball movement direction was analyzed from four lengths of a line. Our proposed method can detect five gaze directions that can be used for selecting four menus on the display monitor. The mean accuracy of our proposed method to detect eye movements for each gaze is 88.6%.
AB - Each country's number of people with disabilities and strokes increases yearly. Hand defects and stroke make them have limitations in doing activities. It caused their hand has paralyzed. Hence, they find it difficult to do daily activities, such as running a wheelchair, choosing a menu on the screen display, and so on. One solution offered is utilizing eye movement as a navigation tool that can replace the role of the user's hand, so they can run a wheelchair independently or choose a menu selection on display by themselves through the movement of their eyes. Detection of eyeball movements in this study only utilizes a camera as a sensor mounted in front of the user. So that it is more practical and easier to use than if we have to pair an electrooculography sensor in the area around the user's eyes. This research proposed a new approach to detect the five gazes (upward, downward, leftward, rightward, and forward) of the eyeball movements by using Backpropagation Neural Network (BPNN) and Dynamic Line Sector Coordinate (DLSC). Line Sector Coordinate is used to detect the eyeball movement based on the pupil coordinate position. The eyeball movement direction was analyzed from four lengths of a line. Our proposed method can detect five gaze directions that can be used for selecting four menus on the display monitor. The mean accuracy of our proposed method to detect eye movements for each gaze is 88.6%.
KW - detection
KW - Disabilities
KW - dynamic line sector coordinate
KW - eyeball movements
KW - eyeball movements
UR - http://www.scopus.com/inward/record.url?scp=85148283777&partnerID=8YFLogxK
U2 - 10.30630/joiv.7.1.1011
DO - 10.30630/joiv.7.1.1011
M3 - 期刊論文
AN - SCOPUS:85148283777
SN - 2549-9904
VL - 7
SP - 37
EP - 43
JO - International Journal on Informatics Visualization
JF - International Journal on Informatics Visualization
IS - 1
ER -