Abstract—The goal of this study is to support the walking of
visually impaired people using AI and a small camera. It is
difficult for the visually impaired to walk straight toward the
target due to the characteristics of the disability. Even if they
think they are walking straight, they will move off to the left and
right. For this reason, they may have an accident such as falling
from the platform of the train. In this report, in order to enable
visually impaired people to walk safely toward the target, a
small camera attached to the body captures the direction of
travel, detects the area where the pedestrian can walk safely
from the recorded pedestrian, and guide users there. We
consider that the place where there are pedestrians is an area
where we can walk safely. Pedestrians are detected from
camera images using AI. Use sound to inform the user of the
direction of travel. We prototyped these methods.
Index Terms—Visually impaired, veering tendency, deep
learning, wearable device
Yasuyuki Murai and Yumiko Ota are with Nihon Pharmaceutical
University, Tokyo 113-0034 Japan. E-mail: arisue-888@nichiyaku.ac.jp
(Y.O.)
Hisayuki Tatsumi and Masahiro Miyakawa are with Tsukuba University
of Technology, Ibaraki 305-8521 Japan. E-mail:
tatsumi@cs.k.tsukuba-tech.ac.jp (H.T.); mamiyaka@gmail.com (M.M.)
*Correspondence: murai@nichiyaku.ac.jp (Y.M.)
[PDF]
Cite: Yasuyuki Murai, Hisayuki Tatsumi, Yumiko Ota, and Masahiro Miyakawa, "Prototype of a Method to Support the Walking of Visually Impaired by Detecting the Walkable Area Using Pedestrians," International Journal of Engineering and Technology vol. 15, no. 2, pp. 41-44, 2023.
