Hacker News Board Hacker News, browsed like a classic message board
Live data from Hacker News

Making Visible Watermarks More Effective

research.googleblog.com

Post reply on HN
Print view View original

31–40 of 98 posts

Re: Making Visible Watermarks More Effective

#31

Post by overcast »

FTA wrote: post #19

This reminds me a bit of the common argument for locks: it keeps regular people honest. Watermarks are designed really to deter you from casually copying an image and pasting it on your site, unless you don't care at all about a watermark showing (e.g. a blog site with a few followers). I'm sure a mixture of the original computer vision technique plus some smearing on the original image could wash away even these ran…

This is exactly right, and is also what keeps the casual thief from doing things like copying software. Obviously these can be cracked, but the layman isn't going to deal with, or even know about it. I don't want to get into the "well those people wouldn't have paid for it anyhow" argument. I firmly believe, as much as a hassle as it is, it's preventing proliferation to the masses. Not much to complain about really, we did it to ourselves.

Same can be applied to locks on a house. Prevents someone looking to have fun, from just walking into your home. Otherwise it's too much effort.

Regarding reporting. I've submitted various copyright claims to Instagram for others using my photography, and they responded quickly. It was a smooth transaction.

Re: Making Visible Watermarks More Effective

#32

Post by AgentME »

jotato wrote: post #12
fake-name wrote: post #7

Earlier quoted context omitted.

The point of the article is that, for sites with a huge image corpus, if they apply some randomization to the watermark for every image, it's much more challenging to recover a un-marked image. Right now, sites like shutterstock and other image sites apply the same watermark to every image, making it quite easy to computationally extract the watermark, and then apply the inverse transform to marked images. If the wat…

I wonder what would happen if I took 1000 copies of the same image that is using a warped mark and ran it through the algorithm....

The algorithm takes a bunch of pictures, figures out what's in common between them, and then removes that part. If you give it only many copies of the same image, it will find that the whole image is in common between the images, and subtract out the entire content.

Re: Making Visible Watermarks More Effective

#33

Post by 0x0 »

jotato wrote: post #27
lightbyte wrote: post #20

Earlier quoted context omitted.

The algorithm is calculating the mean of the set of images, so putting 1000 copies of the same image will do absolutely nothing

i mean to remove the watermark on that specific image

They would probably use the same warp every time on a specific image, so that you can't average it out across multiple download sessions.

Re: Making Visible Watermarks More Effective

#34

Post by azdle »

jotato wrote: post #27
lightbyte wrote: post #20

Earlier quoted context omitted.

The algorithm is calculating the mean of the set of images, so putting 1000 copies of the same image will do absolutely nothing

i mean to remove the watermark on that specific image

The idea is to use a randomly-warped per-image watermark. So in theory, there's exactly one distinct watermark per image, so 1000 of the same image would have the same watermark.

Re: Making Visible Watermarks More Effective

#35

Post by Ajedi32 »

sp332 wrote: post #25
Ajedi32 wrote: post #13

I wonder if removing watermarks from images would be something a machine learning algorithm would be well suited for. Admittedly my understanding of machine learning is rather limited, but it seems like it'd be pretty easy to generate a large training set of watermarked and unwatermarked images for the algorithm to train on. No idea how effective it'd be though.

Not all machine learning is "deep neural networks". I think that's why your comment is being downvoted. The video shows specific model parameters that are learned for each watermark.

Yes, apologies for being imprecise in my wording. In this case I am indeed specifically referring to deep neural networks.

Re: Making Visible Watermarks More Effective

#36

Post by throwaway91111 »

fake-name wrote: post #7
ouid wrote: post #3

Seems like it is saying the following: We can find the watermark in images and subtract it from the image. If we distort the original watermark, but subtract the average watermark, then you will not recover the original image. duh?

The point of the article is that, for sites with a huge image corpus, if they apply some randomization to the watermark for every image, it's much more challenging to recover a un-marked image. Right now, sites like shutterstock and other image sites apply the same watermark to every image, making it quite easy to computationally extract the watermark, and then apply the inverse transform to marked images. If the wat…

Surely the watermark could also be ugly and irreversible; ie clamp to white.

Re: Making Visible Watermarks More Effective

#37

Post by throwaway91111 »

grenoire wrote: post #23
user5994461 wrote: post #8

Why do they use transparent watermark????

It is not so easy to decide if the watermark covers an important aspect of the photo.

Yea, but that may be acceptable for eg previewing a photographer's work—you reach out for the originals.

Re: Making Visible Watermarks More Effective

#38

Post by tyingq »

dtech wrote: post #17

It seems the problem here is that the watermark is transparent, and thus still contains information about the original image. Similar to how if you want to censor a part of an image, you should always use a single solid color because e.g. a blur can be inverted. But I guess that would degrade the quality of the watermarked photo too much.

People do sometimes use small, non transparent watermarks.

If they are small enough though, then the person copying just overlays their own. Or uses automated inpainting to approximate what's underneath.

Re: Making Visible Watermarks More Effective

#40

Post by nxsynonym »

icebraining wrote: post #26
nxsynonym wrote: post #4

Thinking out loud here - but would this be a good use case for blockchain technology? The problem with visible watermarks is it detracts from the image visibility. Nobody wants to look at photos or digital art pieces with huge ugly watermarks on them. Could blockchain tech help establish ownership in a way that would make watermarking obsolete? edit: cool - downvotes for asking a question. Real nice guys.

cool - downvotes for asking a question. Real nice guys. I didn't downvote you, but I suspect at this point people are a bit fed up of the constant suggestions of using the Blockchain for everything.

Ugh - if that's the case this place is less open to discussion than I thought. I was honestly curious if it could be a possible application for a technology that people constantly complain has no use.

Oh well, keep em coming. The downvotes only make me stronger ;)

Post reply on HN