Robust periocular recognition by fusing sparse representations of color and geometry information


We propose a re-weighted elastic net (REN) model for biometric recognition. The new model is applied to data separated into geometric and color spatial components. The geometric information is extracted using a fast cartoon - texture decomposition model based on a dual formulation of the total variation norm allowing us to carry information about the overall geometry of images. Color components are defined using linear and nonlinear color spaces, namely the red-green-blue (RGB), chromaticity-brightness (CB) and hue-saturation-value (HSV). Next, according to a Bayesian fusion-scheme, sparse representations for classification purposes are obtained. The scheme is numerically solved using a gradient projection (GP) algorithm. In the empirical validation of the proposed model, we have chosen the periocular region, which is an emerging trait known for its robustness against low quality data. Our results were obtained in the publicly available FRGC and UBIRIS.v2 data sets and show consistent improvements in recognition effectiveness when compared to related state-of-the-art techniques.

Examples of periocular images of different subjects with and different illumination conditions (FRGC) and varying gazes (UBIRIS.v2), containing the corneal, eyebrows and skin regions.

(a) Periocular images from the FRGC database

(b) Periocular images from the UBIRIS.v2 database

ROC curves for periocular recognition using the FRGC data set. a ROC curves for the proposed REN model with different downsampling ratios. b-c ROC curves of the models implemented by Park et al. [35] and Woodard et al. [50], respectively.


J. C. Moreno, V. B. S. Prasath, G. Santos, H. Proenca. Robust periocular recognition by fusing sparse representations of color and geometry information. Journal of Signal Processing Systems, 82(3), 403-417, Mar 2016. doi:10.1007/s11265-015-1023-3

Preliminary version at arXiv, September 2013, doi:10.48550/arXiv.1309.2752

Related work:

J. C. Moreno, V. B. S. Prasath, H. Proenca. Robust periocular recognition by fusing local to holistic sparse representations. 6th International Conference on Security of Information and Networks (SIN), Aksaray, Turkey. Proc. ACM Digital Library. pp. 160-164, November, 2013. doi:10.1145/2523514.2523540