Ever lasting problem of Fakes 0

[Berserker Art]

Ever lasting problem of Fakes

below are a link or two for fakes on ebay. does anyone have any suggestions as to how we can properly deal with this epidemic, its happening more and more often and bugs me a lot. its clear reporting to ebay does nothing. some of them have been conning buyers for years like (fosland)

 

 

http://www.ebay.co.uk/itm/Jack-Kirby-Daredevil-ink-drawing-on-paper-original-comic-art/262967506800?_trksid=p2045573.c100033.m2042&_trkparms=aid%3D111001%26algo%3DREC.SEED%26ao%3D1%26asc%3D43649%26meid%3Dbef310c4d47048aa8ec79ac3fbf146a5%26pid%3D100033%26rk%3D3%26rkt%3D7%26sd%3D232323542501

 

 

http://www.ebay.co.uk/sch/artforsale01/m.html?item=232323542501&hash=item36178fede5%3Ag%3AKKIAAOSwCQZZC5Aj&rt=nc&_trksid=p2047675.l2562

[suspense39]

http://theartnewspaper.com/market/art-market-news/nowhere-to-hide-new-tool-brings-technical-firepower-to-the-fight-against-fraudsters/

 

[Madman1138]

There are several eBay accounts from Hungary that sell the fakes. I think they are all related or maybe even the same seller on the different accounts, and the main red flags are the typical Dr. Seuss sketches and they really like Bruce Timm a lot. Other than reporting them to eBay, there really isn't anything that can be done. They are ruining it for everyone and people like that need to be driven out, but that still leaves how? How to drive them out of the hobby?

 

Dave K

[thethedew]

7 hours ago, Madman1138 said:

How to drive them out of the hobby?

 

Dave K

A greater volume of complaints from Ebay bidders might help get Ebay's attention.  

(That said, I haven't complained myself, so please don't feel I'm criticizing)

[Taylor G]

Since Ebay doesn't appear to have any interest in doing anything about it, perhaps we could take a "community policing" approach.  For example, this could take the form of a central repository on the Web of these fakes.  It would be indexed by artist.  Going to an artist's page, one could see a collection of fake images recorded for that artist.  The image itself would have to be uploaded by someone, since Ebay auction listings eventually disappear, along with a short description of where it was seen and why it was suspected to be fake.

A logical way to do this would be to extend ComicArtFans.  Allow people to upload these images and not have them show up in their regular CAF galleries, and have these "virtual" galleries of fake images for an artist be filled with images contributed by CAF members.  Use the rating system for images to allow CAF members to provide feedback on images recorded as fake (so a high rating would be "bad").  

Surely a better use of disk space than all the garbage crayon GGA out there.

This could be raw input for machine learning systems that can sort out fake auction listings on Ebay, so eventually potential buyers could just use a browser plugin to detect the fakes.  This would have to be kept open, because ML is going to make mistakes, so honest sellers would need to know why their stuff is labelled fake so they can defend themselves.

Having invested some money in this hobby, I am concerned about its long term viability in the digital age.  It shouldn't be that hard to police these kinds of half-assed fakes, but as these fakes make good money, the hobby will attract better fakes.  Who knows to what extent it's already happening, they're just smart enough that the marks don't know they've been conned.  I'd expect to see a lot of this in the shallow end of the pool, kept under the radar where organizations like Ebay can't be bothered policing it.  

[Unca Ben]

2 hours ago, Taylor G said:

Since Ebay doesn't appear to have any interest in doing anything about it, perhaps we could take a "community policing" approach.  For example, this could take the form of a central repository on the Web of these fakes.  It would be indexed by artist.  Going to an artist's page, one could see a collection of fake images recorded for that artist.  The image itself would have to be uploaded by someone, since Ebay auction listings eventually disappear, along with a short description of where it was seen and why it was suspected to be fake.

A logical way to do this would be to extend ComicArtFans.  Allow people to upload these images and not have them show up in their regular CAF galleries, and have these "virtual" galleries of fake images for an artist be filled with images contributed by CAF members.  Use the rating system for images to allow CAF members to provide feedback on images recorded as fake (so a high rating would be "bad").  

Surely a better use of disk space than all the garbage crayon GGA out there.

This could be raw input for machine learning systems that can sort out fake auction listings on Ebay, so eventually potential buyers could just use a browser plugin to detect the fakes.  This would have to be kept open, because ML is going to make mistakes, so honest sellers would need to know why their stuff is labelled fake so they can defend themselves.

Having invested some money in this hobby, I am concerned about its long term viability in the digital age.  It shouldn't be that hard to police these kinds of half-assed fakes, but as these fakes make good money, the hobby will attract better fakes.  Who knows to what extent it's already happening, they're just smart enough that the marks don't know they've been conned.  I'd expect to see a lot of this in the shallow end of the pool, kept under the radar where organizations like Ebay can't be bothered policing it.  

If it's feasible (and not to imply it isn't) this is a good great idea.

[comicwiz]

Reproductions of valuables has been an ongoing issue in numerous collecting categories. The idea of tackling it always comes squarely down to raising awareness of sellers who demonstrate a pattern of passing off fakes, but this is imperfect as it doesn't guarantee new collectors or people who aren't part of any community won't get ensnared in a deal with a known fraudster.

With paintings, it's a little more complex because there are a number of reproduction "makers" who will reproduce works by hand, or using oil painting printers. They activiely solicit sales through websites and seem to know how much wiggle-room to operate within, using terminology and safe zones where these types of scams can run unabated.

This idea of a software performing regular searches isn't a new one. I'm certain it's derivative of the kind of metasearch search/retrieval software used by stock image companies, although those rely on metadata to be attached to every image they post online. Certainly, a proprietary software stands a better chance of incident discovery if the front line approach is to attach metadata to any/all images posted, and collaborating with artists. The imperfect nature is that social sharing is engineered to reduce file size, which does mean at times any/all metadata gets stripped. This can be either deliberate if certain tools are chosen by more crafty fraudsters, or accidental.  Bottom line, there needs to be a human judgement component, so any software being heralded as an approach to stamp out the problem will become a great way to sell "additional" services to improve reliability and effectiveness.

Where things get tricky is following the trail to the actual perpetrators. Having seen this very thing going on in other hobbies, I can say it isn't easy to put these guys out of business, and the willingness of people to chase down infringers/violators operating in International zones known for this type of thing is extremely difficult, and often times not viable.

[Taylor G]

What I'm suggesting is not searching based on metadata, that sounds pretty useless, it's using image recognition and AI to have a program (rather than a human) say, I've seen that before.  It would be the same kind of technology that is used for facial recognition and categorization of people's holiday photographs.

Another thing one could try, is to use ML to recognize when a piece of art for sale on Ebay is a fake copy, because someone other than the seller has it in their possession.

You could find this out by what's on CAF.  Since not everyone wants their stuff up for public display, this would be done best if people could make their art available for ML without making it available for public display.

Of course, this now provides a compelling reason for someone to take their stuff off CAF once they no longer own it.

Before everyone gets their hopes up that recognition of fakes can be completely automated, it should be recognized that image recognition can be fooled if you know what you're doing, by messing with the image pixels in ways that are not detectable to the human eye.  But:

1. The fakers that have been exposed are not very sophisticated.
2. Perhaps image recognition will get better (we don't really understand what it does now).

The ML stuff is maybe a bit ambitious.  It would be nice to at least have a central repository of images of fakes, indexed by artist, this is not technically hard, it just requires some resources (maybe a Kickstarter?).

I don't know how widespread fakes are, but it is possible it is much more prevalent than people think, and only the amateur hour fakes are being exposed.

What really worries me are Artists Editions, all those high resolution scans of valuable art....

[comicwiz]

1 hour ago, Taylor G said:

What I'm suggesting is not searching based on metadata, that sounds pretty useless, it's using image recognition and AI to have a program (rather than a human) say, I've seen that before.  It would be the same kind of technology that is used for facial recognition and categorization of people's holiday photographs.

As useless as it may sound, I'd take the metadata approach every time over what you suggest.  I say this as a software developer of search retreival software, and I know the analytic side of what you are describing is practically useless to combat the problem unless it's paired with human review. Incident discovery isn't the problem as much as wading through the irrelevant results and the only way to determine those insights are actionable is to assign that duty to a real person with the skill and expertise to determine whether it's even worth pursuing. I built a fraud detection system for the insurance industry almost 10 years ago, and the system took us far enough to locate personal property reappearing on the market after it had been reported as stolen.  One incident I can discuss involved a VIN on a car reported as stolen reappearing with a new ownership VIN record. While the technology worked to a tee, what I soon realized was there was a business side to this endeavour, and the reality-dice learning moment was watching the insurance industry picking and choosing which battle they wanted to fight everytime our system caught something.

[Taylor G]

25 minutes ago, comicwiz said:

I know the analytic side of what you are describing is practically useless to combat the problem unless it's paired with human review. 

We're not in disagreement.  The "machine learning" would come from human feedback to the robot about whether or not it was correct in categorizing something as fake.  

I'm not thinking of Deep Thought adjudicating what is or is not a fake on Ebay.  I'm thinking of a browser plugin that someone points at an Ebay image, and it tells the user whether or not it thinks it's a fake.  If the user asks for an explanation, they get directed to a Web page with matching images from its database of images.  The user can give their feedback there.

This requires that all the images be publicly available, so there's that.

And how to stop the bad guys from misdirecting the software with their feedback?  Voting won't help if the bad guys can rent time on botnets to rig the voting.  But maybe all you need to do is change the economics of the game for the fakers.

[comicwiz]

50 minutes ago, Taylor G said:

I'm not thinking of Deep Thought adjudicating what is or is not a fake on Ebay.  I'm thinking of a browser plugin that someone points at an Ebay image, and it tells the user whether or not it thinks it's a fake.  If the user asks for an explanation, they get directed to a Web page with matching images from its database of images.  The user can give their feedback there.

A browser extension that functions like a "faux" filter would be something I could get behind. Even if it can't be initially as useful due to database ramp-up requirements, depending on the context/composition of the piece, another technique could cross-compare it's production counterpart. Text in word balloons and line art matching is impossible to replicate perfectly. The Daredevil pin-up is something of a misfit in the kind of working framework I could see being achievable, because the faker intentionally used a page appearing to show age (foxing) and using the image from DD #1 or the first page splash for comparison could be subverted with a claim it was an old comission Jack Kirby did.

It goes without saying, but the extension/plugin would probably be more reliably adept at showing comparisons of signatures as they evolved as one factor in the process of determining a fake from real.

Edited May 6, 2017 by comicwiz