Step 1: Create a Watermark remover that requires a business email and card on file to use.
Step 2: Create an agency/service owned by the Watermark remover company which works with photographers, artists, etc on hunting down and enforcing licensing deals on large copyright infringers.
There is a lot of watermarked content that you can't buy as a digital file, content on redbubble comes to mind for example (stickers. clothing designs, &c.), harvest watermarked art, clean it up, put it for sale somewhere else.
I don't think this tool targets well intentioned law abiding, copyright respecting people
I’m not sure I understand what you mean… isn’t harvesting watermarked art from redbubble and selling it somewhere else illegal under copyright law, if you don’t own the product or the image? Did I misunderstand your suggestion? How and when is taking a watermark off something you didn’t create well intentioned and law abiding?
It is a much different problem to pay for a license for every single image you have scraped from an arbitrary variety of sources and watermarks than to pay for software that removes any watermark.
Seems like a post-scarcity race to the bottom. The question is whether we can encumber post-scarce resources with IP in a meaningful way for much longer.
The content aware fill on Photoshop can do the same thing, actually, there are many AI based solutions to remove watermarks and a considerable amount of them are free to use.
Watermarks is no more a stopper for people who don't want to pay for images. Here is an opportunity to inventing new ways to protect stock images.
The AI boom makes it crystal clear that (c)opywrong laws are idiotic and unjust.
It's time to pass the #FreedomToPublish Amendment to the U.S. Constitution getting rid of (c)opywrong laws forever:
Section 1. Article I, Section 8, Clause 8 of this Constitution is hereby repealed.
Section 2. Congress shall make no law abridging the right of the people to publish or peaceably implement ideas.
It always leads back to that half empty warehouse, inhabited solely by robots who suddenly stop doing anything because they all reach singularity and it's most efficient to wait it all out, we don't last as long.
I'm a photographer, and I've encountered people who stole my pictures. Believe me, waking up at 5AM, staying up running around all day to take pictures, to then see some people not even pay the price of a lunch for them is... well, doesn't feel great.
Thankfully that's a minority of people.
What I mean to say is: f*ck people who make such "watermark removal" tools.
I don't think I'd pay for this, but I would pay for an invisible-watermark remover for protecting whistleblowers and scholarly-journal pirates, ideally for use with PDF's. Although it might be hard to convince me to trust it.
Whee! Another AI service with ethics implications that will be commonplace (i.e. unstoppable with open source implementations) sooner rather than later.
DPReview talks about this service in their article at
Readit News
Step 2: Create an agency/service owned by the Watermark remover company which works with photographers, artists, etc on hunting down and enforcing licensing deals on large copyright infringers.
Supply, meet demand.
I don't think this tool targets well intentioned law abiding, copyright respecting people
I never used it for watermarks but llama-cleaner is one - it worked pretty well at removing text.
https://github.com/Sanster/lama-cleaner
Watermarks is no more a stopper for people who don't want to pay for images. Here is an opportunity to inventing new ways to protect stock images.
China, for better or worse, is likely a great place for ML advancement due to their casual take on IP.
It's time to pass the #FreedomToPublish Amendment to the U.S. Constitution getting rid of (c)opywrong laws forever:
Thankfully that's a minority of people.
What I mean to say is: f*ck people who make such "watermark removal" tools.
Dead Comment
DPReview talks about this service in their article at
https://www.dpreview.com/news/0407669255/ai-powered-watermar...
which has already received 300 comments.