Rabitə və İnformasiya Texnologiyaları Nazirliyinin elektron xəbər xidməti
New artificial intelligence recognizes any emotion through a simple video camera
A direct question to the person does not always lead to a truthful answer, because all people tend to dissemble. And the producers of media content would like to know what kind of plot twist or scene in the episode, especially liked the viewer. And they are willing to pay for a system of recognition of the true emotions so much that the proposal was not long in coming – in the company Silver Logic Labs had developed a specialized AI.
Algorithms of emotion recognition by facial expression of a person have significant commercial value and is kept secret. Silver Logic Labs claim that brought the technology to perfection – they just want a digital camera with sufficiently high resolution. She scans the small movements of facial muscles with a frequency of several times per second, which allows not only to recognize the momentary emotion, but also to follow the change of mood of the viewer over time.
Curiously, the hardware requirements for this system of emotion recognition is very modest. For example, it is able to work with wearable police cameras, which allows you to quickly give an answer – whether the suspect is bluffing when he says that he’s unarmed? And if you adapt it for a video surveillance system in a hospital or nursing home, it is possible to predict the beginning of an attack the patient, when a person only making an involuntary grimace slightly from the pain.
Applying technology to security cameras in the crowd, you can try to calculate the suicide bomber with a bomb or a drug addict under the influence of substances. The teacher will understand how to perceive the lesson, the students, the Deputy will read the thoughts of the voters, the lovers sort out their feelings, and diplomats will have a formidable weapon for negotiations. By the way, this is not fiction — the developers report that the system is already almost unmistakably recognize people under the influence of drugs.
23/08/17 Çap et