h a l f b a k e r yNeural Knotwork
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,
|
|
|
Please log in.
Before you can vote, you need to register.
Please log in or create an account.
|
To take the perfect group photo:
1. Film a short video clip of the group standing still.
2. Load that video clip into the software this idea proposes. It might run on the camera itself, or separately on a personal computer.
3. Have face recognition software locate each face in the group;
manually fill in faces the software missed.
4. Either manually or automatically (smile analysis, etc.), move each individual face back and forward in time using a hardware knob or software slider, until finding that moment of the clip that best captures that person's expression.
In other words, you're creating a temporal map that for each pixel says which point in time that pixel should come from. You don't really notice the composition in the result - it merely appears as if everything happened to be just right.
This can be applied to any photograph that shows a complex situation involving slightly changing actors; I'm using group photos as an example because there's already software to automate some of this.
inspired by ...
Better_20group_20photos
which picks one perfect moment doing something in real time - which is much harder. [jutta, Jul 23 2009]
[link]
|
| |
A sensible transition between digital cinematography and photography. |
|
| |
An alternative implementation could work out random fidgit cycles to make a Potterseque "wizard picture". |
|
| |
love it! etch-a-sketch your photograph! |
|
| |
There was that Japanese chap (I think) who made image installations where touching an area caused that area to shift forwards or back in time in a sort of ripple type effect. I remember being memorised by a view over a bridge that extended from evening into daybreak. |
|
| |
//memorized// sp: mesmerized, mesmerised
(unless [zen_tom] was actually in the image...) |
|
| |
Photos are notoriously difficult to compost, because of the
platicised gloss finish and all the dyes. |
|
| |
Beat me to it. That was what I kept misreading in my eye's mind. |
|
| |
My experiments with composites run into trouble with both viewpoint and target angle variations. |
|
| |
Reminds me of a long-ago shot in National Geographic, taken by the camera to be used for the Mars Viking missions. The caption said something about the people at the left of the photo going home several hours before the ones at the right were imaged. |
|
| |
As long as the photo is stored as a film along with the temporal map to create the picture i like the idea. If the film is cut down to the synthetic picture, it would better serve to have everyone bring 3d-surface-mapped scans + internal models of themselves, and simply render the whole scenery in optimal lighting, optimal facial expressions and matched clothing. |
|
| |
You could add some worms to your temporal photo compositing bin to speed the process. |
|
| |
This sounds similar to what is already being done with sound recordings. If a musician hits a bad note, they replace it with a note copied from a different part of the recording, or a different recording. If a face has an ugly expression, replace it with a different expression from a different part of the video. I don't see any reason why it wouldn't work. Bravo! |
|
| |
I suspect this kind of thing is already being done, but probably with multiple stills instead of video, and with tedious Photoshopping. You just want to make the process more efficient with specialized software. That still makes it an original idea, in my opinion. |
|
| |
I actually do this quite a bit. When I need a good photo, it's easier to hit record for a few seconds and choose the right frame later, and masking takes care of getting individual faces right. But +1 for the software that does it automatically. |
|
| |
What is the definition of 'perfect' ?, real ? |
|
| |