Computer Science – Performance
Scientific paper
Dec 2005
adsabs.harvard.edu/cgi-bin/nph-data_query?bibcode=2005aspc..347..168r&link_type=abstract
Astronomical Data Analysis Software and Systems XIV ASP Conference Series, Vol. 347, Proceedings of the Conference held 24-27 Oc
Computer Science
Performance
Scientific paper
Image analysis, such as component fitting of radio interferometric images has traditionally been based on likelihood techniques applied to deconvolved images. The analysis usually ignores uncertainties arising from fitting flux components to extended emission as well as from the process of deconvolution itself. One would therefore like to estimate the properties of components representing the entire emission present in the raw, dirty image. In practice, this is not feasible given the large dimensionality of the parameter space. We present a Bayesian approach in which we fit elliptical Gaussian components to sub-regions of the dirty image, taking into account the point spread function. Our method samples the posterior distribution to estimate the relative probabilities and uncertainties associated with the number of flux components and their parameters. This information can be used to augment the process of object detection and characterization. We compare the performance of this approach to the standard methods.
Cornwell Tim J.
Rau Urvashi
No associations
LandOfFree
Monte Carlo Image Analysis in Radio Interferometry MC-FIT: A Bayesian Approach to Object Detection does not yet have a rating. At this time, there are no reviews or comments for this scientific paper.
If you have personal experience with Monte Carlo Image Analysis in Radio Interferometry MC-FIT: A Bayesian Approach to Object Detection, we encourage you to share that experience with our LandOfFree.com community. Your opinion is very important and Monte Carlo Image Analysis in Radio Interferometry MC-FIT: A Bayesian Approach to Object Detection will most certainly appreciate the feedback.
Profile ID: LFWR-SCP-O-1034512