TY - UNPB
T1 - Predicting Face Recognition Performance Using Image Quality
AU - Dutta, Abhishek
AU - Veldhuis, Raymond
AU - Spreeuwers, Luuk
N1 - Submitted to TPAMI journal on Apr. 22, 2015. Decision of "Revise and resubmit as new" received on Sep. 10, 2015. At present, updating the paper to address the feedback and concerns of the two reviewers. The re-submitted paper will be uploaded as version 2 on arXiv
PY - 2015/10/24
Y1 - 2015/10/24
N2 - This paper proposes a data driven model to predict the performance of a face recognition system based on image quality features. We model the relationship between image quality features (e.g. pose, illumination, etc.) and recognition performance measures using a probability density function. To address the issue of limited nature of practical training data inherent in most data driven models, we have developed a Bayesian approach to model the distribution of recognition performance measures in small regions of the quality space. Since the model is based solely on image quality features, it can predict performance even before the actual recognition has taken place. We evaluate the performance predictive capabilities of the proposed model for six face recognition systems (two commercial and four open source) operating on three independent data sets: MultiPIE, FRGC and CAS-PEAL. Our results show that the proposed model can accurately predict performance using an accurate and unbiased Image Quality Assessor (IQA). Furthermore, our experiments highlight the impact of the unaccounted quality space -- the image quality features not considered by IQA -- in contributing to performance prediction errors.
AB - This paper proposes a data driven model to predict the performance of a face recognition system based on image quality features. We model the relationship between image quality features (e.g. pose, illumination, etc.) and recognition performance measures using a probability density function. To address the issue of limited nature of practical training data inherent in most data driven models, we have developed a Bayesian approach to model the distribution of recognition performance measures in small regions of the quality space. Since the model is based solely on image quality features, it can predict performance even before the actual recognition has taken place. We evaluate the performance predictive capabilities of the proposed model for six face recognition systems (two commercial and four open source) operating on three independent data sets: MultiPIE, FRGC and CAS-PEAL. Our results show that the proposed model can accurately predict performance using an accurate and unbiased Image Quality Assessor (IQA). Furthermore, our experiments highlight the impact of the unaccounted quality space -- the image quality features not considered by IQA -- in contributing to performance prediction errors.
KW - cs.CV
M3 - Preprint
BT - Predicting Face Recognition Performance Using Image Quality
ER -