New on LowEndTalk? Please Register and read our Community Rules.
All new Registrations are manually reviewed and approved, so a short delay after registration may occur before your account becomes active.
All new Registrations are manually reviewed and approved, so a short delay after registration may occur before your account becomes active.
Comments
OP mentioned he scraps this content so yeah
Content on imageboards usually lasts a couple of days before it's purged and lost forever. I'm archiving the posts and making it available for an undefined time. The content is already out of there and free to view and download, i just try to organize, tag it and give some metadata (first time the image was reposted, how many time it was reposted, where are most reposts...).
Many websites do that and are seen as positive from the community, because they make content that would be otherwise be lost forever easy to browse and see.
https://archived.moe/
https://4archive.org/
Artists are also usually happy about their stuff being shared and seen by as much people as possible, fact it that some might not be happy and they could resort to DCMA (99% invalid) takedowns and similar stuff even before contacting me.
I thought i was clear in my OP
Right, but do you take a consent from the copyright owner before you re-distribute this work?
Terms on imageboards usually say that when you upload stuff you pretty much lose any right on it. Aka: upload it and you give us and everybody who browses our site rights to do whatever they wish with your stuff.
So you have a contract allowing you to collect this copyrighted material and distribute it with each and every author or other entity that has rights to it ?
have you considered not hosting the site, this may be the easiest solution
Suicide might be the easiest solution to the pain that is life but i'm not suiciding yet
Providing that it is indeed an invalid abuse claim and the rights to the image are not owned by the person reporting then I don't see why you should have a problem with any host. Of course you also need to be careful of pictures depicting minors in many jurisdictions and possibly bestiality laws in others.
Most hosts will get annoyed with you if you are receiving copious amounts of legitimate abuse - yes. But copious amounts of fraudulent abuse should be fine. I know we have procedures for situations like this (normally an attempt at takedown if an attack fails).
OP seems to be running automated crawlers similar to what Google, Archive.org and others all do, which (despite some objections here and there) is de facto accepted as doing public good, as long as they deal with legitimate complaints when received. While there are certainly subtleties involved, I doubt anyone is likely to be persuaded of anything here.
Pictures depicting minors and bestiality are deleted immediately. I can detect photos from the spectrum and frequency distribution of a pic and I put those in a quarantine.
I'm worried that some host would scroll through my website, see the content and go like "wtf is this shit" and immediately terminate me.
Yeah. People who aren't in the "field" just assume i'm doing some shady stuff while it's widely accepted in the community.
@edfox Wait, so you are running some kind of crawler and it deletes any images found except for cartoons, which it stores? How do you know which one is cartoon? Do you recognize by the total number of colors or grayscales being used? How do you differ photographed cartoon drawn on white paper canvas from a completely digital one? Do you apply some small simplification/low entropy/pixel average filter first, then compare as said?
Furry porn? Weird hobbies out there
I use a mix of different methods. Usually an FFT on the image is enough to understand if it's a photo or a digital drawing. Digital drawings don't have sensor noise, have usually a solid background and incredibly saturated colors.
I also check the filename: a DSC_0019.JPG means it's probably a photograph, a tumblr_[hex]_720.jpg means it's a tumblr thumb, photo-2018-02-13-12-18.jpg means it's from Telegram, 10129218.artist-picture.png means it's from Furaffinity, _u18chan.jpg it's from u18chan, a [timestamp]s.jpg it's from 4chan...
I also look at the resolution. Pictures mostly have standard sizes (8mp, 12mp, 16mp...) and ratio (16:9/4:3) while digital drawings are pretty much irregular. I also run an ocr to recognize artists names or other watermarks.
I think Edfox already answered your question: "I can detect photos from the spectrum and frequency distribution of a pic and I put those in a quarantine." A quick web search shows technology for that has been around at least since the 1990s. I don't see a claim that OP's system is 100% reliable or that it automatically deletes anything. It sounds like the photo detection stuff flags some suspicious images for manual review. Stuff left over might trigger some complaints that get handled. I don't think anyone has claimed more than that here.
Images are compared against a database with known valid reliable fingerprints (e621, furaffinity). If they're in the db - they're shown. If not, they're hidden until somebody approves them.
It's not hard
Are you anti-bestiality, out of curiosity edfox
Yes. Though completely friendly with feral drawn images.