Relaxed Simultaneous Tomographic Reconstruction and Segmentation with Class Priors for Poisson Noise

Research output: Book/ReportReportResearch

385 Downloads (Pure)

Abstract

This work is a continuation of work on algorithms for simultaneous reconstruction and segmentation. In our previous work we developed an algorithm for data with Gaussian noise, and in that algorithm the coefficient matrix for the system is explicitly store. We improve this algorithm in two ways: our new algorithm can handle Poisson noise in the data, and it can solve much larger problems since it does not store the matrix. We formulate this algorithm and test it on artificial test problems. Our results show that the algorithm performs well, and that we are able to produce reconstructions and segmentations with small errors.
Original languageEnglish
Place of PublicationKgs. Lyngby
PublisherTechnical University of Denmark
Number of pages19
Publication statusPublished - 2015
SeriesDTU Compute Technical Report-2015
Number6
ISSN1601-2321

Fingerprint

Dive into the research topics of 'Relaxed Simultaneous Tomographic Reconstruction and Segmentation with Class Priors for Poisson Noise'. Together they form a unique fingerprint.

Cite this