The fake painting supervision is costly and time consuming. After the art history experts send the paintings to be supervised to the laboratory, they may need to undergo complex techniques such as gas chromatography, infrared detection, radiocarbon dating, etc. to identify the authenticity. . Now researchers have developed an artificial intelligence (AI) system that analyzes the strokes to determine whether the work is the master of the artist.
According to reports, in the mid-1990s, historian Maurits Michel van Dantzig developed a method of identifying painters by brush or pencil strokes. He found that the length, direction, shape and force of strokes would be unique. Stroke signature."
Recently, Rutgers University and the Dutch Painting Restoration and Research Studio (ARRS) scholars published an AI monitoring system, inspired by van Dantzig's method, which relies on analyzing the brushstrokes of paintings to distinguish true. Pseudo.
The study digitized 297 line drawings of Picasso, MaTIsse, Modigliani and other masters of painting, allowing the system to dismantle 80,000 strokes from it. Then, the Recurrent Neural Network (RNN) is used to learn the characteristics of the strokes to identify the painter.
In addition, the researchers also trained a machine learning calculus system to identify specific features in the stroke, such as the shape of a line in a stroke. The combination of the two systems accurately identifies the artist's chances of reaching 80%. In order to test the effect of the system to identify the product, the researchers also hired the painter to paint in the same style. As a result, the probability of successful recognition by the system was 100%.
In addition to the breakthrough in the art of art supervision, another important contribution is that by comparing the machine learning system trained by the researcher and the RNN two systems, one can always see it as a "black box" (black). Box) How the RNN works. Black box means that the researcher cannot master the operation of a system, and there is no way to explain how the result of the system comes from.
For example, the RNN system for identifying fake products is discriminated by the pressure changes exerted by the painter on the strokes, that is, the force they write. However, the research team also pointed out that this method can only be used to identify paintings with clear lines. In other words, works that are not obvious in terms of strokes are not lost.
The next step for the research team was to use this test method in Impressionist paintings, as well as other brush-finish 19th-century paintings, for further testing.
In many electronic communication equipment and instruments where always use some cable assemblies to connect some electronic units and circuits.
We use different connectors, such as SMB, SMA, N TNC, MCX, IPEX, BNC with different cables such as Flexible cable, semi-flexible cable, semi-rigid cable, corrugated cable etc. Our products widely used in customers' special technology, equipment, skilled operation , advanced test
We test all the finished cable assemblies one by one, Insulation resistance , Dielectric Withstanding Voltage and VSWR at the frequency range form 0~18 GHz. to meet clients' requirements and Characteristic.
RF Connector Coaxial Cable,RF Cable Assembly,SMA Cable Assemblies,N Type Cable Assemblies
Xi'an KNT Scien-tech Co., Ltd , https://www.honorconnector.com