Webhose Uses Image Recognition to Identify Illicit Content on the Dark Web | Hacker News Reader

Webhose Uses Image Recognition to Identify Illicit Content on the Dark Web

webhose.io

22 pointsrangeva9 comments

"The Dark Web is infamous for the trading of illegal weapons around the world."

[citation needed]

yes you can buy guns though mostly it's 3D printed shit where with every shot fired the chance that it blows up in your hand increases. There are some guns which are better quality but they are very probably tainted.

The darknet is lending itself extremely poorly for this type of transaction. Anonymity is exactly what I do not want when obtaining a weapon (yes even you tend to commit a crime with it). you may remain anonymous but the potential history of the gun certainly won't be and is traceable more accurately than any fingerprint.

The only reason why you might want to sell a gun there is to get rid of potential evidence while at the same time pin the crime on another person.

According to a study conducted by RAND Corporation, most firearm vendors on the darknet (59%) originated from the United States, but the combined revenue of European sales are 5 times higher than those of the US.

Most guns sold in Europe are coming via the South-West into Europe, usually from Russia, via Moldova, Serbia, Albania, etc. Small firearms are highly available in Europe if you know where to look. I very much doubt that the sale of all guns on the darknet is higher than 1% from all total trade. It's neither profitable nor practical to sell arms on the darknet.

> The darknet is lending itself extremely poorly for this type of transaction. Anonymity is exactly what I do not want when obtaining a weapon (yes even you tend to commit a crime with it). you may remain anonymous but the potential history of the gun certainly won't be and is traceable more accurately than any fingerprint

Sure, they can trace the gun...but who cares? All they can do is trace it back to the guy who sold it to you. If you're intending to commit a crime with it, I don't see how that's problematic for you.

It's a problem if you haven't done anything (yet) but the previous owner did.

You’re buying a weapon of unknown history in a manner that (in theory, at least) provides no link to you. In what scenarios do you see yourself then interacting with law enforcement to a degree that the history of the weapon would come up but you wouldn’t have otherwise been in trouble?

They search my home or me and find the weapon. If it's clean, I can face punishment for a criminal possession of a weapon. But if a crime was committed with it, I can also be charged with it and I may have a hard time trying to prove it wasn't me.

"Through image recognition and the extraction of meta-data, Webhose was able to identify the GPS location of the image"

How does image recognition help with this? The gps data is right there in the JSON.

Also at the beginning of the article we see:

"Since users want to remain anonymous and criminals often use their own jargon, it is far more difficult to navigate and track the plans of cyber criminals."

So it is very handy that the JSON contains the uncensored urls, which saves us part of this trouble... not very clever.

I believe it's a combined process, i.e image recognition is for finding objects and meta data extraction is to extract GPS location. I don't think the links in the JSON image will help "the criminals" :)

What's the point of sharing the GPS coordinates and the street view pictures of the address? To attract LEA customers? To make others bring the drug vendors to justice? It's not your job to play police here..

That said, the article can be reduced to:

- they are using trained networks to classify images and make those tags searchable

- they are reading exif data from images and make that searchable

That's it. If it's a photo of pills hosted by some hidden services it's “illicit content” per their approach.

the Webhose.io Image Analytics can be used in so many ways.. Detect/Monitor people via GPS location Actors signatures Brand Logo monitoring and more ... will be happy to answer questions.

Reply on news.ycombinator.com