On 2003.10.07 21:23 John Hebert wrote: > At 20:21 10/7/2003 -0500, you wrote: > >On 2003.10.07 17:03 Dustin Puryear wrote: > > > Freddie from the 19'th District Court will not be able to make the meeting > > > tomorrow, and he was our primary speaker. ... Give me your thoughts. > > > > > > >Will Hill thinks, "bummer". > > > >Does anyone know anything about image manipulation coding, specifically > >recognizing areas of focus in an image? I'd like to take a series of > >pictures of an object with depth using an optical microscope and then > >merge the areas of focus into a single image similar to an electron > >microscopy image. Who says you can't get depth of focus optically? My > >brain does it, so should my computer. > > I'm a total amateur here when it comes to this subject, but I guess that > you can't easily detect depth of focus. That would seem to require the > computer to understand metaphor, which is currently kinda hard to do. > > You may be able to mathematically detect "blurriness" or "fuzziness" in the > areas of out of focus, though I haven't seen the image in question. >
Yeah, I'm an amatuer too =;), having never written an app that deals with images. I talked to the author of freeUSP (http://www.freeusp.org/) once about this and he enjoyed the idea and thought that it might be doable. I asked him if contrast between pixels could be used to detect focus. He thought that would work and that his image routines could do it. A serries of images would be taken with different planes of focus. The lowest plane would form the base image and any those areas in each image that score highest in focus would be added on top. I'd rather use GPL'd routines than freeUSP and think that I can find them. I'll bet someone else has already thought of this and done it because it seems so obvious. If not, it's a nice project that can be thrown together from GPL'd code.
