PROJECT TITLE :

Tiled-block image reconstruction by waveletBased, parallel filtered back-projection - 2016

ABSTRACT:

We have a tendency to demonstrate an algorithm, relevant to tomography sensor systems, to obtain images from the parallel reconstruction of essentially localized elements at different scales. This is often achieved by combining methodology to reconstruct pictures from limited and/or truncated knowledge, with the time-frequency capabilities of the wavelet rework. Multiscale, furthermore time-frequency, localization properties of the separable 2-dimensional wavelet transform are exploited as an approach for faster reconstruction. The speedup is realized not only by reducing the computation load on one processor, but additionally by achieving the parallel reconstruction of several tiled blocks. With tiled-block image reconstruction by wavelet-based mostly, parallel filtered back-projection (FBP), we tend to measure a lot of than 36 times gain in speed, compared with standard FBP.