Interesting article. Thanks for posting it.

I would imagine the first camera with this technology is going to be a bit expensive.

What other ways are there (if this isn't off topic) to protect your photos? Being someone who loves to take photos, it would be nice to know how to protect them. If they be of my grandson or of something else.

What's the point of having an expensive system like this onboard a camera, when the photographer can watermark his own photos on his PC before publishing them?

Hi Achernar

How do you go about watermarking your photos? If you wouldn't mind explaining how to do so?

There are programs for that, but none that I've tested myself (I've no use for that).

The root problem/solution here is to come with a standardize solution to watermark images with the ID of the owner. Something like pgp signatures for emails.
But once a system is known, it can be circumvented (watermark tempered, becoming useless).

You can do it using a visible watermark, e.g put your logo on it somewhere. If you really want to protect it use a large logo covering entire image, set a decent transparency and use a gradient fill. That's really hard to remove but of course also ruins the image, only really useful if your trying to give samples.

The ohter way is to use an "invisible" watermark, This encodes it into the image data itself without any visible alterations. Not sure how effective it is because I'd image resampling the image would destroy or partially destroy it. Don't know because I never had one to try and manipulate. :

Watermarks are put into the least-significant bits of the data, in such a way that they are not visible. As far as the eye is concerned, they are random noise.

You MIGHT be able to remove a watermark by altering color depth. Technically, you'd need to chop-off n least-significant bits (set them to to 0 or 1 or probably best, randomize them) but that might visibly degrade the image.

You don't know which pixels contain the watermark data (not all pixels are affected). Only the watermark-owner holds the key that determines which pixels contain the watermark.

Watermarks are designed to survive common image manipulations. There are redundancies, mathematical integrity checks, etc.

The iris system is a huge improvement.
It is a more compelling argument that the owner of the iris watermark is the first one to watermark.

A PC watermark can face a challenge where someone who watermarked an image later might say they took the picture.

Even is someone successfully removes an iris watermark, manipulating it and creating a new iris watermark would be prohibitive and you'd be able to compare the original to the modified and win an argument.

Does this presages the need to submit iris scans to search engines with DMCA notices?

Published iris scans: sounds like doors opening for identity theft.

Out of the thousands and thousands of commercial photos I take each year, I rarely, if ever, have my eyeball on the view finder when I shoot the photo. Sounds annoying.

Seems like a privacy issue, if your retina scan is retained to be used in all your images, could it for instance be a biometric signature that could be used to tag images that you didn't take? Or other things... I have will ask my local resource on that subject.

Like Loner I very seldom look through the viewfinder to take a digital photo.

I don't think this will be desirable.

How accurate are iris scans anyway, if you wear glasses? Mine are almost always dark when I go outside (due to photochromatic lenses), which is where I take the most photos. And like others, I'm not always looking through the viewfinder in any case.

It's not a technology I would be willing to pay extra for, if the choice were between a camera with or without this ability.

What about the new trend for Live View photos taken via the LCD and not the viewfinder? I expect more and more cameras to take on this approach. One day the viewfinder might even disappear. (My current compact doesn't have one!)

What about photos taken using a remote? Or a self-timer?

There are too many situations where your iris wouldn't be at the viewfinder to make this plausible.

Futile, there are enough ways to prove you made something. The real trick is to know when somebody steals your property, and that's not covered by marking it any way.

Oh, and cryptography has plenty of way to digitally sign something (even with the power of the law behind it in some countries). Theres no need for bimetrics inhere.

Iris is accurate to over 1 in 1.2 million, much more accurate than most of the biometric technologies. Generally it is one of the most expensive techs as the people who developed the algorithm charge on a per user license basis.

[edited by: IanTurner at 7:48 pm (utc) on Feb. 25, 2008]

There are too many situations where your iris wouldn't be at the viewfinder to make this plausible.

That's were my issue with the potential of using stored scans to insert into the image will become a problem.

Watermarking might help to identify theft. But you'd need a spider of your own to be out searching the web for that watermark. As someone said, watermarking might work if it were clear and visible. It would require cloning to remove it. And that might degrade the photo, even enough to make it unsuitable for commercial use. But an invisible watermark might simply be destroyed by resaving to jpg, or something, or running through various subtle filters.

The first trick is a) know your photo has been stolen. It's a big web. There are lots of sites.

B) prove it was yours. The best way would be to have the original. It can be digital. But it should contain more detail than what is on the web. So, ideally, any photo you publish is a slight crop of the original. If you shoot RAW, you have the 'raw'. That might be difficult to 'reverse engineer' from a jpg. It might be that you shot a number of similar photographs. You'd have all of those. The copyist would not. All of that could prove that you were the original photographer.