Confidence measure for temporal registration of recurrent non-uniform samples
PReMI'07 Proceedings of the 2nd international conference on Pattern recognition and machine intelligence
Choice of low resolution sample sets for efficient super-resolution signal reconstruction
Journal of Visual Communication and Image Representation
Hi-index | 0.00 |
This paper presents a method for the reconstruction of bandlimited functions from non-uniform samples by means of synthesis functions. We state explicit worst-case estimates which depend on the bandwidth and the sampling density. The proposed reconstruction method uses oversampling to deduce well-localized synthesis functions which improve the locality properties of the reconstruction. As a consequence, the resulting algorithm is suitable for local recovery of bandlimited signals from sufficiently dense local sampling information. The theoretical results are confirmed by numerical examples based on a finite-dimensional model described.