 |
 | Home | Contact Us | FAQ | Search & Site Map | Link to Us |
 | Sign In | Join | Other 45 Sites in Network |
Natural Science Forum / Physics / Optics / June 2007Deconvolution Advertising
Hi All, I just would like to comment on an advertisement. Biophotonics ran an ad for MediaCybernetics AutoQuantX deconvolution process. I commend MediaCybernetics for their efforts to mainstream the application of deconvolution and blind deconvolution. These methods have proved their worth in countless published papers and deserve wider application. I have objection with the "64-bit deconvolution" bubble. I believe this is a statement that their software uses 64-bit processing, however it is misleading. Since most imagers are at most 12-bits, you will not produce 64-bits of information after deconvolution (although you may get 16). There is no indication that this will improve your results. The bit depth of your imager, and the algorithm you use, are much more important than the number of digits your computer holds while processing. Best Regards, Jim C >Hi All, > [quoted text clipped - 16 lines] >Best Regards, >Jim C The improvement is in speed of results. Autoquant's blind 2D deconvolution is excrutiatingly SLOW. A 64-bit approach allows for 32-bit floating numbers to be processed more quickly and it breaks the 4GB addressing limit of 32-bits. This lets the app have more RAM and thus reduce paging to/from disk. A 64-bit OS also needs 64-bit hardware platform to make this work. Kiss French. Drink California. gary at gaugler dot com Gary, Anything you can do to speed up image processing methods is appreciated. Bit depth is an important issue in deconvolution. Performing deconvolutions on images with higher bit-depth gives you a much better shot at an effective restoration. Perhaps its just my perspective, but I thought the ad gave the impression that you could derive 64-bits from either the image or point spread function. Perhaps "64-bit processing would have been better." Jim C > The improvement is in speed of results. Autoquant's blind 2D > deconvolution is excrutiatingly SLOW. A 64-bit approach allows for > 32-bit floating numbers to be processed more quickly and it breaks the > 4GB addressing limit of 32-bits. This lets the app have more RAM and > thus reduce paging to/from disk. A 64-bit OS also needs 64-bit > hardware platform to make this work. >Gary, > [quoted text clipped - 14 lines] >> thus reduce paging to/from disk. A 64-bit OS also needs 64-bit >> hardware platform to make this work. Not sure where they talk about this 64-bit topic. The app will handle 8-bit RGB and greyscale as well as 12-bit, 14-bit and 16-bit TIFF. The < 16-bit words are zero-filled from MSB down to the highest valid data bit. I rarely used it for RGB LM pix. My main use was greyscale SEM TIFF. These files can be very large as 8-bit and really big as 16-bit (extended 14-bit data). Running 10-15 iterations would take all day and likely wind up crashing. Small RGB files from low rez digicams go pretty fast. But it still takes a lot of putzing to get a decent result. Their LM flourescence example appears to me as more of a brightness and contrast issue. I'd run PS first and then Lucis Pro. I guess I'm still a dealer for Autoquant but the cost of the app is so high ($4,500) that there is almost no viable interest in it. For hard core LM work at low to medium rez, that is different. But for SEM, I use and sell Lucis Pro. For LM I would use Fovea plug-in for PS. Kiss French. Drink California. gary at gaugler dot com > >Gary, > [quoted text clipped - 27 lines] > > gary at gaugler dot com Gary, You may want to check out my stuff at www.quarktet.com/Gallery1.html. In addition to what is shown (application to X-ray and microscopy images), I am working on a paper on applying blind deconvolution to TEM imaging, to be presented in November. Jim C Jim C |
Reply to this Message |
 Sign In
 Join
 My Latest Posts My Monitored Threads My Blog My Photo Gallery My Profile My Homepage Start New Thread Enable EMail Alerts Rate this Thread
|
©2009 Advenet LLC Privacy Policy - Terms of Use
This website includes both content owned or controlled by Advenet as well as content owned or controlled by third parties.