There is no lost data between the pixels of a digital image or the grains of a negative. You can't recover what was never there to get lost. I recall reading an article in an IEEE journal about a method for removing motion blur. You could change the problem to removing camera shake instead of dealing with a lack of focus. I also assume it might be possible to correct for aborations in the lens. If you have a reference object near the markings of interest you might be able to correct for some focus problems. Think of how giant telescopes use a laser beam as a reference and deform the mirror to correct for any apparent smearing of that reference. You might be able to use the knife edge as a reference. Assume the edge is straight or actual a nice smooth curve. Apply appropriate (this might be too much like magic) transformation to clear up the knife edge. Since the markings are very close to the edge they might get cleaned up too. Iv'e never heard of this type of image improvement, but sounds plausable. Tim Corio On Fri, 2007-04-13 at 17:17 -0700, Robert Earnest wrote: > A friend of mine is writing a book and one of the plot twists involves using > technology to "recover?" informatin from a 4x6 prints found at a crime > scene. > > Here is what she wrote. I had an idea but I thought it would be really fuun > to pose this challenge to you guys. Any ideas, anyone? > > >From her letter to me... > > ______________________________________________________ > > > INFO: > > > A 4x6 snapshot, left at a crime scene, plays an important role in the plot > of the book I'm writing. The book is set in Seattle. > > A tech at the SPD Police Photo Lab enlarges the snapshot. > > When the Protagonist-Detective sees the enlargement, he notices that are > some markings on a knife in the photo. The markings are too small, however, > to make out. They could be symbols, initials, or simply decoration. > > > NOTE: At this point, I can go for one of two scenarios: > > The first scenario is simple, though not as interesting. I would not need a > scientist's input for it though > > The second scenario is more complex, but also more interesting, > 'technology-wise'. > (NOTE: I can only use it though IF it's true that NASA scientists have > developed a superior method of enhancing photos, and they have advanced > software for doing this. I need to know whether it's true that their > software can 'fill in' (or extrapolate) information that *isn't actually > there*. > > NOTE: > In either scenario, I want to avoid writing a laughable CSI-type scene in > which the CSI tech 'magically' enlarges or enhances a very poor quality > photo (i.e. taken with a camera-phone or low end point-and shoot), to the > point where they are suddenly able to identify an object or person in the > photo ‹ even though the object or person wasn't in focus in the first > place. > > > > > > SIMPLE SCENARIO: > (Photography-wise, this scenario only works if the markings are 'in focus', > but too tiny to make out.) > > The Protagonist-Detective sees the enlargement and notices there are > markings on a knife in the picture. For plot reasons, I don't want the > Detective to go back to the Police Lab to have the photo enlarged. Instead > (because it's the weekend), he takes the snapshot to a friend who is > computer /photography savvy (i.e. Photoshop whiz). She enlarges the snapshot > again and he sees that the markings on the knife are initials. > (Note: It's only on TV and in movies that Police Lab Techs are available > 24/7.) :) > > > > COMPLEX SCENARIO: > > The SPD Police Photo Lab Tech enlarges the snapshot. When the > Protagonist-Detective sees the enlargement, he notices that there are > markings on a knife in the picture. > > The Police Photo Lab Tech tells the Detective that if he enlarges the photo > any further, it will begin 'losing detail rather than gaining it'. He > points out that the markings are 'out of focus' and that it's impossible for > anyone to either enlarge or enhance a a photo and magically obtain a 'clear' > image ‹ if the image is out of focus to begin with, or if part of the > information isn't there. He cites a 'laughable' TV episode in which a CSI > tech did that. He then says he has heard though that NASA scientists have > been able to do this, by using algorithms, and advanced software to > extrapolate the information. (i.e. Hubble Telescope distortions? Other > space photography advancements?) > > The frustrated Detective has a friend, who is a scientist at Microsoft in > Redmond/Seattle. The friend works for NASA in conjunction with Microsoft. > (I've been told that this is viable, that Microsoft and NASA do work > together on various projects.) The friend is familiar with the technology > NASA uses to enhance space photos, and she has access to the advanced > software. The friend 'enhances' the photo for the Detective and they can see > that the markings are initials. > > > COMMENTS? ADVICE? > > Thanks ya'll. > > > R > >
