BROWSE
Detail
Conference Paper
An Annotated Dataset of Human Gait Images for Kinematic Analysis
Cited - time in 
Share 
Share
- Authors
- Jang-Hee Yoo, Byung-Tae Chun
- Issue Date
- 2015-06
- Citation
- Conference on Computer Vision and Pattern Recognition Workshops (CVPRW) 2015 - The Future of Datasets in Vision, pp.1-2
- Publisher
- IEEE
- Language
- English
- Type
- Conference Paper
- Abstract
- We describe an annotated dataset of human gait images, based on the SOTON large population database. To analyze and classify human gait motion, 10 body points (mostly joint positions) are manually extracted from each image of the gait sequences by using an annotation tool. The annotated dataset includes 100 different subjects with seven gait sequences for each subject, a total of 700 gait sequences. In experiments, the quality of the annotated dataset was evaluated by analyzing the forward displacement. Also, angular kinematics described by rotation angles around the joints over time were analyzed using medical data.
- KSP Keywords
- Gait motion, Large population, Over time, Population database, annotation tool, gait sequences, human gait, kinematic analysis, medical data, rotation angle