Abstract
In the previous chapter, the concept of separating data into an evaluation and development dataset for optimization and testing was discussed. In Chaps. 4 through 6 we saw the role a development dataset can play in optimizing biometric system components for best recognition performance. Yet, optimizing a biometric system to a development dataset may expose it to inherent data bias , and development results may lead to the selection of biometric systems that perform poorly in practice. To be certain, a particular biometric system configuration is suitable for practical use it should be put through an evaluation process, whereby it is tested against previously unseen data. As part of our demonstrative GRF-based gait recognition experiment, this chapter puts the best biometric configurations discovered in earlier chapters to the test by applying them to the previously unseen samples from an evaluation dataset.
Access this chapter
Tax calculation will be finalised at checkout
Purchases are for personal use only
Reference
Cattin, Philippe C. 2002. Biometric authentication system using human gait. Ph.D. Thesis, Swiss Federal Institute of Technology, Zurich, Switzerland.
Author information
Authors and Affiliations
Corresponding author
Rights and permissions
Copyright information
© 2016 Springer International Publishing Switzerland
About this chapter
Cite this chapter
Mason, J.E., Traoré, I., Woungang, I. (2016). Measured Performance. In: Machine Learning Techniques for Gait Biometric Recognition. Springer, Cham. https://doi.org/10.1007/978-3-319-29088-1_8
Download citation
DOI: https://doi.org/10.1007/978-3-319-29088-1_8
Published:
Publisher Name: Springer, Cham
Print ISBN: 978-3-319-29086-7
Online ISBN: 978-3-319-29088-1
eBook Packages: EngineeringEngineering (R0)