Remove watermarks from PDFs/papers with pdfparanoia | Hacker News Reader

Remove watermarks from PDFs/papers with pdfparanoia

github.com

26 pointskanzure10 comments

What is the motivation behind this?

This is from the github page: "pdf watermark removal library for academic papers"

I read that, but why? I've never been part of the magical world of academia, and don't understand why watermark removal is something that's important.

Author here (btw, like you I am not a part of the world of academia either). These watermarks include your name, institution, and ip address on every page. This is a huge privacy violation, especially when sharing research with colleagues.

http://scholar.google.com/scholar?q=%22Authorized+licensed+u...

http://scholar.google.com/scholar?q=Redistribution+subject+t...

http://scholar.google.com/scholar?q=Downloaded+from+http%3A%...

http://scholar.google.com/scholar?q=Downloaded+*+*+2001..201...

More details: https://groups.google.com/group/science-liberation-front/bro...

Even more details: https://groups.google.com/d/msg/diybio/o6irBNeCmrE/pqGJXMlQY...

Also, Google seems to have deals with academic publishers based on ip address. You can sometimes see Google in the watermarks in their own search index:

http://scholar.google.com/scholar?q=filetype:pdf+google-inde...

Even stranger, Google Docs Viewer seems to remove watermarks from some pdfs, but I have no idea why. Did they write their own watermark removal library?

evidence: https://groups.google.com/group/science-liberation-front/t/a...

That's awesome! Thanks for explaining this.

It might be cool if you could expand the readme to include this information. "Why?" was the first question that came to mind, perhaps only because I don't have the necessary context or background to understand the utility of this utility at first glance.

I think more likely that Google Docs Viewer just doesn't support vertical text and so it happens to remove the watermark.

I presume this violates your license with the paper vendor?

There are many complicated licensing relationships in this ecosystem.. do you mean a particular ToS? a college's license with Elsevier/JSTOR/IEEE? a scholar's use of library resources, presumably governed by some honor code or something setup by a university? or something else?

Is there anything like this for stock images? There's a simple way(masking) to do it with photoshop- but it's not automatic.

It seems like the producers of stock photography have a library to do this; there's a button on each stock photography page that will do it for you, but I can't find source code for it. You just have to enter a sixteen-digit string and a MM/YY verification for it to work. Reports are that certain 15-digit strings will also work.

Reply on news.ycombinator.com