innovator

Add idea


Calendar

«    November 2018    »
MonTueWedThuFriSatSun
 
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
 

 

Advert

 

Payment

 

Advert

 

Authorization

Стартап

DateDate: 20-11-2018, 06:01

  
 Personal non-property and property rights for the invention belong to the author and protected under current legislation.
 The idea is that in the usual escalator, where we all usually go with standing, there is another parallel path with chairs, that would not be exactly as on the pictures. That is, we will be able now not only to ride standing up, but sitting. Of course, it is very convenient for elderly people who find it hard to stand even on the escalator, mothers with children, and just regular visitors of the metro, if the chairs of the escalator go empty. They have a safe mechanism of exit from the mine of the new escalator, can turn around 90 degrees for more convenient use. 
 The idea is open for investment and complete its foreclosure. 
+380505238948

DateDate: 20-11-2018, 05:56

Fingerprint scanning systems are widely used in modern smartphones and other gadgets - it is believed that they are able to provide a high level of protection for the device. Researchers at New York University have denied this claim. They taught the neural network to forge “living keys” of users and circumvent biometric identification systems.
Neural network has learned to crack fingerprint scanners
In most cases, the sensors do not read the entire surface of the finger, limited only to areas that tightly touch the working surface of the scanner. Scientists have suggested that the number of partial matches among the resulting skin patterns can be quite large.
The DeepMasterPrints algorithm was taught to compare ready-made prints, and then instructed it to generate fake ones on its own. As a result, the neural network was able to bypass the protection system, significantly reducing its effectiveness.
Neural network has learned to crack fingerprint scanners
Previously, a biometric sensor could erroneously react to someone else’s imprint only once in a thousand. The algorithm, who has become accustomed to the “fingers” forgery, has learned to deceive the sensor in just five attempts.
Moreover, he repeatedly managed to replace the original with a self-made copy, even without having information about the original.
Neural network has learned to crack fingerprint scanners
Earlier, Counterpoint Research published a study, according to which more than 50 percent of smartphones released in 2017 are equipped with fingerprint scanners. By the end of 2018, according to experts, their number will increase by another 20 percent. And although it’s still too early to worry about data integrity in your device, manufacturers should consider strengthening the protection of their gadgets today.