Affects of footwear clothing etc on gait recognition

Members do not see these Ads. Sign Up.

Hi

It's quite possible in the excitement of challenging SALRE that this might be missed but hey!

There are all kinds of issues to do with the affect of clothing etc (co-variate factors) on recognising people when attempting gait recognition. The following conclusion from a recent paper by the folks at Southampton might be of interest. At least to those with an interest.

Conclusions
"In this chapter, we have investigated the impact of the different covariate factors on the performance of gait recognition using kinematic-related features. Four different covariates are analysed including footwear, load carriage, clothing and walking speed.

Based on a covariate-based probe dataset of 440 samples, a high recognition rate of
73.4% is achieved using the KNN classifier with k = 5. This is to conclude that people identification using dynamic gait features is still perceivable with better recognition rate even under the different covariate factors. The footwear, clothing and load carriage covariates are observed to have almost no effects on the performance of gait recognition with similar results when walking with barefeet or without carrying bags. However, the gait recognition drops largely when walking with flip flops or wearing a trench coat due the difficulties encountered during the extraction of dynamic gait features using the model-based method."

With regards to this latter sentence I suspect that the trained human eye may fair a little better as it recognition of gait cues is not predetermined.

Ian

 

Hi Ian,

If I ever feel like holding up my local bank at least I know what to wear now ;-)

Looks like an interesting read - do you have the full reference for the article?

Thanks

 

Hi Ian

It is a small paper but quite helpful and was released 2008. It is done in their "Let's see if it can be done lab" and built around computer gait recognition - way above my head.

Of course there is the co-variable issues of different surfaces but in the years I've done this I have been surprised at the ability to possibly recognise people when we might think we should not.

Ian

Exploratory Factor Analysis of Gait Recognition

Imed Bouchrika & Mark S. Nixon
Department of Electronics and Computer Science
University of Southampton, SO17 1BJ, UK
fib04r, msng@ecs.soton.ac.uk