IRIS super-resolution via nonparametric over-complete dictionary learning

Aljadaany, Raied; Luu, Khoa; Venugopalan, Shreyas; Savvides, Marios

doi:10.1109/icip.2015.7351527

Cited by 6 publications

(3 citation statements)

References 14 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…Sparse representation in over-complete dictionaries was used in the work of Aljadaany et al [40]. Traditional approaches in this regard, such as K-Singular Value Decomposition (K-SVD), have the limitation that the number of dictionary items and the number of sparse coefficients has to be predefined.…”

Section: Learning-based Methods In the Pixel Domainmentioning

confidence: 99%

A Survey of Super-Resolution in Iris Biometrics With Evaluation of Dictionary-Learning

et al. 2019

View full text Add to dashboard Cite

The lack of resolution has a negative impact on the performance of image-based biometrics. While many generic super-resolution methods have been proposed to restore low-resolution images, they usually aim to enhance their visual appearance. However, an overall visual enhancement of biometric images does not necessarily correlate with a better recognition performance. Reconstruction approaches thus need to incorporate the specific information from the target biometric modality to effectively improve recognition performance. This paper presents a comprehensive survey of iris super-resolution approaches proposed in the literature. We have also adapted an eigen-patches' reconstruction method based on the principal component analysis eigen-transformation of local image patches. The structure of the iris is exploited by building a patch-position-dependent dictionary. In addition, image patches are restored separately, having their own reconstruction weights. This allows the solution to be locally optimized, helping to preserve local information. To evaluate the algorithm, we degraded the high-resolution images from the CASIA Interval V3 database. Different restorations were considered, with 15 × 15 pixels being the smallest resolution evaluated. To the best of our knowledge, this is the smallest resolutions employed in the literature. The experimental framework is complemented with six publicly available iris comparators that were used to carry out biometric verification and identification experiments. The experimental results show that the proposed method significantly outperforms both the bilinear and bicubic interpolations at a very low resolution. The performance of a number of comparators attains an impressive equal error rate as low as 5% and a Top-1 accuracy of 77%-84% when considering the iris images of only 15 × 15 pixels. These results clearly demonstrate the benefit of using trained super-resolution techniques to improve the quality of iris images prior to matching.

show abstract

Section: Learning-based Methods In the Pixel Domainmentioning

confidence: 99%

A Survey of Super-Resolution in Iris Biometrics With Evaluation of Dictionary-Learning

et al. 2019

View full text Add to dashboard Cite

show abstract

“…where the parameter µ y (µ ỹ) is the average gray value of y (ỹ), the parameter σ y (σ ỹ) is the variance of the gray values of y (ỹ), and σ yỹ is the covariance of y and ỹ. By default, c 1 = (0.01 * 255) 2 and c 2 = (0.03 * 255) 2 [79]. Also, the window size is of 11×11, which is weighted with a circular Gaussian filter of standard deviation 1.5 before calculating local statistics.…”

Section: Performance Metricsmentioning

confidence: 99%

“…Regarding learning-based methods, several algorithms have been proposed to learn the mapping between low-and high-resolution images, for example Multi-Layer Perceptrons [70], Markov networks [51], or Bayesian modeling [2]. Some works have also proposed to super-resolve images in the feature space, instead of the pixel domain.…”

Section: Iris Super-resolutionmentioning

confidence: 99%