[CSBi-events] Mark Your Calendar

[csbi-events@mit.edu]

High Performance Deconvolution and Effects on Downstream Image Analysis
Hans van der Voort, Ph.D.
CEO & CSO, Scientific Volume Imaging The Netherlands

10.30 am Thursday September 29th  McGovern Auditorium, Whitehead
Contact: James G. Evans (jgevans at wi.mit.edu)


Abstract:
Deconvolution aims to recover the 'true' object distribution from an 
image distorted by, among others, diffraction and noise. The 
resolution measured in terms of the Half Intensity Width (HIW) in 
deconvolved images can be much higher than in the original, under 
ideal conditions near-isotropic sub-100nm HIW values can be achieved. 
However, in confocal or two-photon fluorescence microscopy the 
available signal strengths do not allow such extremely high HIW 
values, but there deconvolution can significantly improve the axial 
resolution, while effectively reducing noise.

The introduction in recent years of fast Nipkow disk based confocal 
microscopes has made it possible to acquire large 3D-time series. On 
top of that, the number of voxels per 3D image frame is steadily 
increasing. Since deconvolution is generally a compute intensive 
procedure these two ongoing developments result in a rapidly growing 
demand for high performance deconvolution systems.

  In this talk we will discuss high performance deconvolution 
algorithms and their effect on resolution. In addition the effect of 
deconvolution on the reliability of co-localization coefficients will 
be discussed in detail.


-- 


Linda K. Earle, Outreach Coordinator
Computational and Systems Biology (CSBi)
Phone: (617) 324-0074
Fax: (617) 324-0081

Massachusetts Institute of Technology
77 Massachusetts Avenue
Building 68 - Room 459
Cambridge, MA 02139