DissabledGait: Gait Dataset of Normal People and People with Disabilities

Published: 28 May 2024| Version 2 | DOI: 10.17632/v6hy35ydch.2


The DissabledGait dataset provides a collection of walking videos of people with disabilities and normal people. people with disabilities are divided into 2 groups, namely with assistive devices and without assistive devices. From these videos, image extraction, annotation and labeling are performed. There are 130 videos and 6500 images that have been labeled. In the .txt label there are 3 classes indicated by the first number in the file. The number 0, indicates assisstive, number 1 indicates non-assisstive and number 2 indicates normal. Assistive here is a person who has a physical disability in the legs and uses a walking aid in the form of a cane, not a wheelchair. Whereas non-assistive means a person who has a physical disability of the legs, but can still walk. Normal people are people who do not have physical disabilities and can walk like most people.


Steps to reproduce

DissabledGait was taken using a 12MP Redmi note 5 camera. The video taken was then edited by censoring the face area. The edited video results were then cut into several videos. After that, the video is extracted into 50 image frames. The image is then annotated using roboflow to generate a label.


Universitas Nusantara PGRI Kediri, Universitas Negeri Malang


Disability, Gait, Instance Segmentation