SUM Researchers about to Launch Emotion Recognition System

An invention of young scientists from the “Digital Economy and High Technologies” Project Laboratory of the State University of Management (SUM DEHT PL), based on the artificial neural network, helps recognise human emotions and emotional changes. SUM will be launching a trial version of the system in October.

The recognition technology is based on complex mathematical algorithms: a computer analyses almost one billion different parameters of a human face. Numerous photographs are used as the basis for recognition.

‘First off, we need to recognise the face, followed by the position of the face. To work effectively, the system needs a huge amount of data; at least 100,000 photos representing each type of expression: sad, happy and neutral (when the person is calm). We use two separate sets of images: a general one, which can be accessed by anyone, and our own set with photos of undergraduates and staff members,’ says Mikhail Sutulov, Head of the SUM DEHT laboratory.

The number of recognisable emotions is gradually growing, and researchers are also planning to add behavioural analysis (gait, postures, types of movement) to the system before rolling out the final product.

 ‘We can market this system to businesses, for example, to help sellers cater to customer needs: they can start offering tailored products to certain clients, based on the emotions detected by the neural network. Moreover, this technology is an integral part of security systems,’ commented Mikhail Nachevsky, Head of the SUM Project Laboratory.

Scientists at the SUM lab are also working on the following projects: “MoskvaPomosch” (“Moscow Help”)—a digital bulletin board where Moscow residents can solicit or offer help to each other; “Digital Paper”—a new way of storing and transmitting large amounts of hard copy data using a unique encrypting algorithm; “SUM ChatBot”—a virtual online assistant for students, professors and employees of the State University of Management; “YurTech”—a platform with elements of machine learning that will be instrumental in creating algorithms for detecting correlations in Russian laws; and others.

<p><iframe src=”https://www.youtube.com/embed/_q4D5hBzcbs” width=”100%” height=”400″ frameborder=”0″ allowfullscreen=”allowfullscreen”></iframe></p>

Подпишись на тг-канал "Наш ГУУ"