ken@turtlevax.UUCP (Ken Turkowski) (07/13/85)
In article <305@ur-laser.uucp> nitin@ur-laser.uucp (Nitin Sampat) writes: > at what point do you start separating computer > graphics from a field called image processing ? It all depends on the computational model: +----------+ \ +-------------+ | Computer |---->| Frame Store | Computer Graphics (CG) +----------+ / +-------------+ \ +----------+ +--->| Computer |-----+ | / +----------+ | | | Image Processing (IP) | +-------------+ / | +---| Frame Store |<--+ +-------------+ \ Of course, variants are allowed, such as a source and target framestores, which may be virtual. There is cross-breeding, though, to the benefit of both disciplines: Texture mapping is an image processing technique that is used in to render realistic computer graphic images. Curve fitting Pattern recognition (perhaps a discipline unto itself, but more closely related to image processing than computer graphics) has been used to create a geometric model, which is then visualized using computer graphic techniques. Examples of these are noise reduction in CAT scans and such, and font representation by polygons or splines. -- Ken Turkowski @ CADLINC, Menlo Park, CA UUCP: {amd,decwrl,hplabs,nsc,seismo,spar}!turtlevax!ken ARPA: turtlevax!ken@DECWRL.ARPA
tcs@usna.UUCP (Terry Slattery <tcs@usna>) (07/15/85)
> In article <305@ur-laser.uucp> nitin@ur-laser.uucp (Nitin Sampat) writes: > > at what point do you start separating computer > > graphics from a field called image processing ? I prefer the quote made by Dr. Azriel Rosenfeld a few years ago on the topic of computer graphics/processing/imaging: The world can be broken into four pieces; the first has nothing to do with images. The second is generally called computer graphics and deals with generating visible images from numbers (display a line drawn from 0,0 to 1,1). The third, image processing, deals with generating an image from another image, generally with certain enhancements. The fourth, computer vision, deals with generating numbers from images (reading the numbers from bank checks). So, we get: 1. numbers -> computer -> image (show me the line) 2. image -> computer -> image (anti-alias this image) 3. image -> computer -> numbers (what was that check number?) For the net.graphics group, I think enough overlap in various areas exist that a good number of the people will be subscribing to all sub-groups. -tcs Terry Slattery U.S. Naval Academy 301-267-4413 ARPA: tcs@brl-bmd UUCP: decvax!brl-bmd!usna!tcs
nather@utastro.UUCP (Ed Nather) (07/16/85)
> I prefer the quote made by Dr. Azriel Rosenfeld a few years ago on the > topic of computer graphics/processing/imaging: > > The world can be broken into four pieces; the first has nothing to > do with images. The second is generally called computer graphics > and deals with generating visible images from numbers (display a line > drawn from 0,0 to 1,1). The third, image processing, deals with > generating an image from another image, generally with certain > enhancements. The fourth, computer vision, deals with generating > numbers from images (reading the numbers from bank checks). > > So, we get: > 1. numbers -> computer -> image (show me the line) > 2. image -> computer -> image (anti-alias this image) > 3. image -> computer -> numbers (what was that check number?) > > For the net.graphics group, I think enough overlap in various areas > exist that a good number of the people will be subscribing to all > sub-groups. > Terry Slattery U.S. Naval Academy 301-267-4413 Well, this categorization totally misses the operations astronomers refer to as image processing, where we are concerned with extracting quantitative values from digital images -- values which represent, in some way, the physics involved in the processes the images represent. I suspect other disciplines do similar, but different, things. I find the proposed categories too limited to be of much value. -- Ed Nather Astronomy Dept, U of Texas @ Austin {allegra,ihnp4}!{noao,ut-sally}!utastro!nather nather%utastro.UTEXAS@ut-sally.ARPA
dmt@mtgzz.UUCP (d.m.tutelman) (07/18/85)
> > .....So, we get: > > 1. numbers -> computer -> image (show me the line) > > 2. image -> computer -> image (anti-alias this image) > > 3. image -> computer -> numbers (what was that check number?) > > > > For the net.graphics group, I think enough overlap in various areas > > exist that a good number of the people will be subscribing to all > > sub-groups. > > Well, this categorization totally misses the operations astronomers refer > to as image processing, where we are concerned with extracting quantitative > values from digital images -- values which represent, in some way, the > physics involved in the processes the images represent. I suspect other > disciplines do similar, but different, things. I find the proposed > categories too limited to be of much value. Am I missing something? It looks like a nice categorization to me, if you don't mistake the example for the scope of the category. The astronomical problem is clearly (to me, at least) image->computer->number, or IMAGE ANALYSIS. OK, so it isn't ONE number. But the essence of image analysis is to go from the pixels to the ABSTRACTION of which they are an image. Reading a check for text is only one such example. Finding a person in a photo is another [harder] example. But if you read "number" to mean "semantics of the abstraction", this is a much better system of nomenclature than the one I proposed a few days ago. Dave Tutelman Physical - AT&T Information Systems Holmdel, NJ 07733 Logical - ...ihnp4!mtuxo!mtgzz!dmt Audible - (201)-834-2895