h a l f b a k e r yThe halfway house for at-risk ideas
add, search, annotate, link, view, overview, recent, by name, random
news, help, about, links, report a problem
browse anonymously,
or get an account
and write.
register,
|
|
|
Back in the dawn of time you used to
have
to write down the aperture and shutter
speed you had used for your photos if
you
were interested in capturing that kind of
information. Then came a set of high-
tech
cameras which either printed the
information on unused space between
the
frames
(e.g. the Pentax 67 II) or held it in
camera memory to download
later.
Digital camera make this
even
easier by storing more information than
you could possibly ever need in the EXIF
file header but you've still got to go and
look for this. I was struck recently that
most photos I take seem to have some
writing in them somewhere (the name on
the crossbar of my son's bike, the
lettering
on the side of an aeroplane, some text
on
someone's t-shirt, etc.) and the exposure
informtion would be a lot more
acccessible if the camera software could
seek out areas of text in the image and
replace the text with exposure
information, having carefully matched
the
font from a vast library. So, you'd see
photos of people wearing, not shirts with
"Cardinals" across the front, but "f/4 @
125", shop awnings would say "f/8 @
250"
in the same font usually used for "Pepsi",
etc.
The Optics Pages
http://www.vanwalree.com/optics.html
"...A collection of illustrated articles on the chief causes of image degradation in photography." [bristolz, Nov 23 2004]
Please log in.
If you're not logged in,
you can see what this page
looks like, but you will
not be able to add anything.
Annotation:
|
| |
I like it, but something tells me it's not going to be particularly easy. |
|
| |
Now correct me if I'm being dense, but aren't the exposure details only really useful for when you're developing the photograph? And if so, you wouldn't be able to read them until after the point at which they would have been worth knowing - Wouldn't this be like, I don't know, like writing the address on the inside of an envelope? |
|
| |
Actually, [zen], they are useful as a reference for enlargement printing & as feedback for later photographs (at least for me). I think this idea has some merit, as long as the function can be switched off. I would hate to have text that was intentionally in the photograph obscured.
Also, I can see how this can lead to a "where's wally?" game in your photos, trying to figure where it appears in pictures that dont have prominent text. What happens if you are doing portraits & there's no text in view? |
|
| |
Yikes. Finding and matching the lighting, perspective, dof, fov, color, aberrations, and font; building a patch to remove the original information and then performing a composite is much, much more than OCR at work; it's full fledged semantics and then some. I think that at some point in the future this is plausible but for at least the next 10 years . . . I'm skeptical. |
|
| |
Hmm. The patch building (auto-cloning) is already happening in certain packages (MS Digital Image 10, for example) but it is far from perfect and especially if the surrounding element isn't big enough to get a good clone source. I do expect, though, that in a few years one might be able to remove unwanted objects from an image in-camera and perhaps even at the time the shot is taken. |
|
| |
The camera knows about it's lens settings and perhaps, if so equipped, could also build a 3D data image with IR or sonar to yield decent scene surface orientation data for matching the optics and perspective, up to some reasonable short distance from the camera. |
|
| |
The OCR, though, is just plain tough and especially with the type of far flung and custom fonts that turn up in average snapshots (not to mention the perspective, skew distortion, etc.). |
|
| |
The display face library would be huge (and expensive to license) and many of those faces on outdoor advertising and t-shirts are one-off hand drawn fonts. The font library would be necessary because it's unlikely that the characters you are replacing would contain the all of needed letters you need for the replacement phrases. |
|
| |
[bristolz] That's pretty much what I would have said if I'd had the energy. And the knowledge. Ahem. |
|
| |
It would be trivialy easy for a camera to add the desired text automatically in a border (that could be cropped out or not). |
|
| |
[bris] You'll notice that I didn't say that
it'd be easy. Or feasible or sensible, now
I come to think of it... |
|
| |