PROJECT TITLE :

Signal-Level Information Fusion for Less Constrained Iris Recognition Using Sparse-Error Low Rank Matrix Factorization

ABSTRACT:

Iris recognition systems operating in less constrained environments with the subject at-a-distance and on-the-move suffer from the noise and degradations within the iris captures. These noise and degradations significantly deteriorate iris recognition performance. During this paper, we have a tendency to propose a unique signal-level information fusion method to mitigate the influence of noise and degradations for fewer constrained iris recognition systems. The proposed method is predicated on low rank approximation (LRA). Given multiple noisy captures of the same eye, we tend to assume that: one) the potential noiseless images lie during a low rank subspace and 2) the noise is spatially sparse. Based on these assumptions, we obtain an LRA of noisy captures to separate the noiseless images and noise for information fusion. Specifically, we tend to propose a sparse-error low rank matrix factorization model to perform LRA, decomposing the noisy captures into a coffee rank element and a sparse error part. The low rank part estimates the potential noiseless pictures, while the error element models the noise. Then, the low rank and error components are utilized to perform signal-level fusion separately, manufacturing two individually fused pictures. Finally, we tend to combine the 2 fused images at the code level to produce one iris code as the ultimate fusion result. Experiments on benchmark information sets demonstrate that the proposed signal-level fusion technique is ready to achieve a generally improved iris recognition performance in less constrained setting, compared with the present iris recognition algorithms, especially for the iris captures with serious noise and low quality.