Russian physicists and physiologists have developed a technique for recognition of things people imagine, by means of electroencephalography. NeuroG technology was created by researchers from faculty of physics of Moscow State University and the Institute of Higher Nervous Activity and Neurophysiology of Russian academy of sciences.
The technique so base upon rapidly developing technology “Brain-computer interface”, which provides control for external technical devices by means of electric signals of brain. Such systems are often used in game industry or in medicine for rehabilitation of neurological patients.
Researchers based their work on experiments, which were conducted during recent 10 years and involved functional magnetic resonance imaging (fMRI). These experiments demonstrated that parts of human brain, responsible for imagining different objects, have different locations inside a brain. Instead of expensive magnetic resonance imagers researchers used affordable electroencephalographs.
The idea is that every image, which appears in human brain (a sandwich or warm sea), has its own signature electroencephalogram. When a database of electroencephalograms of various images is created, we can use it for image recognition with possible applications in biometric systems, psychoanalysis and etc.
In order to create an extensive database for the project, researchers opened a website, where a participant in the experiment could remotely transfer information, imported by sensors from his brain. For this purpose a participant should put on a small wireless device on his head and connect it to a computer by means of a wireless connection. Electric signals from brain travel to a remote server via a special interface.
An extensive database of electroencephalograms, linked to images and keywords, will allow its creators to understand whether human beings are capable of mastering such a technology and to adapt classifiers for as many people as possible. Currently researchers investigate how many object classes the interface is able to recognize, and can it recognize individual objects, belonging to the same class. Further studies will be focused on finding out how many zones there are in human brains, which are responsible for imagining groups of objects, and how do they interact. Researchers also want to know whether it is possible to unify images of abstract things, like “politics” or “innovations”.
The technology has a wide range of applications, like cyber-search, when a human being is recognized as an image and found via a searching engine like Google.
Source: Science & Life
Kizilova Anna