Crowd-sourced amputee gait data
Crowd-sourced videos of bilateral transfemoral amputee gait were analyzed and the results were compared to known values in previous studies. The results showed that this might be a viable method of collecting data from a wider range of people.
Gait lab research has assisted in expanding our understanding of factors that influence amputee gait. However, recruitment can be challenging for participants to attend lab sessions. Crowd sourcing allows the general public to perform the data collection and allow analysis by the scientific community.
Youtube was searched for existing videos uploaded by the public. The videos were analysed frame by frame, to include 5 complete gait cycles. This was then used to calculate other gait parameters and compared to existing data from previous literature.
The results showed some areas of significant difference compared to the literature: intact stance time, gait cycle time and prosthetic stance phase. Although this was not consistent across all compared literature and a number of areas of analysis were found not to be significantly different to the literature.
It is surprising that there is so much similarity between the results recorded here and in laboratory conditions, suggesting that natural gait is not far removed from that performed in laboratory testing. However, it must be remembered that there is no ‘gold standard’ of results with which to compare the results in this study and fully validate.
Read an interesting article about the pros and cons of crowdsourced data here!
What else could you use crowd source videos to analyse?
> From: Gardiner et al., PLoS One 10 (2016) (Epub ahead of print). All rights reserved to The Author(s). Click here for the Pubmed summary.