Colloquium Computing Science



Date:                          Monday, February 11 2008         

Speaker:                   L.J. van Vliet, TU Delft      

Room and time:      5161.0267 (Bernoulliborg), 16.00




Title:   Robust Super-Resolution: Algorithms and Performance Measures.



Super-resolution restoration addresses the problem of restoring a high-resolution scene from multiple degraded low-resolution images under motion. Due to imaging blur and noise, this problem is ill-posed. We propose two strategies to tackle this problem. The first employs a robust fusion step based on structure-adaptive normalized convolution. The second solves an inverse problem without explicit additional smoothness constraints to obtain a stable solution. While adding a regularization term to the cost function is a standard practice in image restoration, we propose a restoration algorithm that does not require this extra regularization term. The robustness of the algorithm is achieved by a Gaussian-weighted L2-norm in the data misfit term that effectively ignores intensity outliers. With the outliers suppressed, our solution behaves similarly to a maximum-likelihood solution in the
presence of Gaussian noise. The effectiveness of our algorithm is demonstrated with super-resolution restoration of real infrared image sequences under severe aliasing and intensity outliers.

Colloquiumcoördinator is Prof.dr. M. Aiello, Dr. M. Biehl (e-mail: