Showing papers by "Nalini K. Ratha published in 2011"

Secure and Robust Iris Recognition Using Random Projections and Sparse Representations

Abstract: Noncontact biometrics such as face and iris have additional benefits over contact-based biometrics such as fingerprint and hand geometry. However, three important challenges need to be addressed in a noncontact biometrics-based authentication system: ability to handle unconstrained acquisition, robust and accurate matching, and privacy enhancement without compromising security. In this paper, we propose a unified framework based on random projections and sparse representations, that can simultaneously address all three issues mentioned above in relation to iris biometrics. Our proposed quality measure can handle segmentation errors and a wide variety of possible artifacts during iris acquisition. We demonstrate how the proposed approach can be easily extended to handle alignment variations and recognition from iris videos, resulting in a robust and accurate system. The proposed approach includes enhancements to privacy and security by providing ways to create cancelable iris templates. Results on public data sets show significant benefits of the proposed approach.

Automatic Fingerprint Recognition Systems

Abstract: An authoritative survey of intelligent fingerprint-recognition concepts, technology, and systems is given. Editors and contributors are the leading researchers and applied R&D developers of this personal identification (biometric security) topic and technology. Biometrics and pattern recognition researchers and professionals will find the book an indispensable resource for current knowledge and technology in the field.

Practical computer vision: example techniques and challenges

Abstract: Humans, as well as many living organisms, are gifted with the power of "seeing" and Bunderstanding[ the environment around them using their eyes. The ease with which humans process and understand the visual world is very deceiving and often prompts us to underestimate the effort and methods needed to build practical, effective, and inexpensive computer vision systems. In essence, humans have a 500-million-year head start due to evolution; it is extremely difficult at this point to build a computer vision system that has the abilities of a three-year-old child. However, by confining ourselves to particular domains, we can often find shortcuts to solve particular problems. This paper illustrates a number of such solutions in various areas developed by our group at IBM. These include object finding for video surveillance, person identification via biometrics, inspection of manufactured items along railways, and scene understanding for driver assistance, as well as object recognition and motion interpretation for retail stores. We discuss the real-world constraints for each system and describe how we overcame the irksome variability inherent in each task. By further analyzing such successful systems and comparing them to each other, we can come to understand the common underlying problems and thus start to extend our initially limited areas of competence into a more general-purpose vision toolkit. This paper concludes with a set of challenging unresolved problems that if solved could spur great progress in practical computer vision.

Biometric score fusion through discriminative training

Abstract: In the multibiometric systems, various matcher/modality scores are fused together to provide better performance than the individual matcher scores. In [1] the authors have proposed a likelihood ratio test (LRT) based fusion technique for the biometric verification task that outperformed several other classifiers. They model the genuine and the imposter densities by the finite Gaussian mixture models (GMM, a generative model) whose parameters are estimated using the maximum likelihood (ML) criteria. Lately, the discriminative training methods and models have been shown to provide additional accuracy gains over the generative models, in multiple applications such as the speech recognition, verification and text analytics[5, 7]. These gains are based on the fact that the discriminative models are able to partially compensate for the unavoidable mismatch, which is always present between the specified statistical model (GMM in this case) and the true distribution of the data which is unknown. In this paper, we propose to use a discriminative method to estimate the GMM density parameters using the maximum accept and reject (MARS) criteria[8]. The test results using the proposed method on the NIST-BSSRI multimodal dataset indicate improved verification performance over a very competitive maximum likelihood (ML) trained system proposed in [1].