Dear Xuejun

>the software but the hardware and window95. More importantly, we have the
>setup together for both data acquisition and processing and the setup is
>used heavily for data acquisition, we are thinking of getting a more
>powerful standalone image processing workstation.

We are using Bitplane's (http://www.bitplane.ch) Imaris 3D visualization
package on  a SGI O2  for 3D reconstructions of our confocal data sets
(from a Leica TCS).

Recently we demoed the Huygens System from Scientific Volume Imaging
(http://www.svi.nl/), which can be used as standalone deconvolution
software package or as an integrated module of our Bitplane 3D software. We
do not have a "turn key" deconvolution system where microscope, camera and
computer are integrated and therefore where looking for standalone
software, running on our SGI Unix hardware.

I can not say  much about performance because we did not test our data sets
on other systems, but deconvolution of a typical data set of 100 images,
256x256 pixels, 8-bit, took about 30min on our hardware (SGI O2, R5000,
512MB RAM). Roughly, an R5000 processor is comparable to a Pentium Pro on
the PC platform. Although, we had the feeling that our SGI workstation
(http://www.sgi.com) was the lower limit for hardware. You would probably
want to look into a dual processor SGI Octane with 1 GB of memory to run
the Huygens system.

What impressed me most with the Huygens system was the excellent
documentation of the software that takes you step by step through the
process of generating point spread functions (PSF) and the  different
deconvolution algorithms. Also, there are several options for generating
PSFs from beads or directly from your image data sets.

Furthermore, the PSFs can be calculated differently depending on the
characteristics of your optical image acquisiton system, i.e. widefield,
confocal, 2PLSM or 4pi.

I think, the package accepts several image formats including general tiff
image stacks and the Imais file format.

Clearly the price tag is not small. I guess the sofware package is in the
10k range and the necessary hardware can easliy go over 20k.

Best,
Tilmann

Tilmann M. Brotz
mailto:[log in to unmask]
UC Berkeley, Dept. Mol. & Cell Biology
221 Life Sciences Addition  #3200
Berkeley, CA 94720-3200
+1(510)643 9594 FAX: +1(510)643 5624