A computer program to detect fake paintings


Two Picasso paintings


A research team from Rutgers University in the United States and Atelier for painting restoration and research in the Netherlands have developed a computer program that can identify counterfeit paintings.
The researchers found that brush strokes have distinctive characteristics that can not be repeated. They programmed a computer to recognize the painter's pattern and differentiate between original and fake paintings with a success rate of up to 100%.
The researchers disassembled 300 written drawings by Picasso, Matisse, Modigliani, Shell and other artists to 80,000 single brush strokes.
An artificial intelligence system called a deep recurrent neural network was used to learn how to identify each artist's brush strokes.
Researchers also used an algorithm to look for specific features, such as a line shape in a brush stroke.
The researchers said the computer system was able to detect counterfeit paintings every time they were shown.

If you found this post helpful on Steemit, would you please --upvote-- it and ''follow me'' because you will then be able to see more posts like this in your home feed
follow me

Coin Marketplace

STEEM 0.21
TRX 0.20
JST 0.034
BTC 90362.36
ETH 3112.51
USDT 1.00
SBD 2.96