Saratov State Technical University scientists Gagarin Yu.A. developed a human emotion classifier based on neural networks designed to assist professionals in the medical, psychology, marketing and entertainment industries. This was reported to socialbites.ca by the Russian Ministry of Education and Science.
This development is based on neural networks technology that can simulate the functioning of the brain. They are used for image recognition, speech, image processing and other tasks where high accuracy and speed are required.
“Research shows that emotions play an important role in human life. Classifying and understanding them can help with a variety of issues related to mental health, relationships with other people, and even performance at work. We automate this process using modern methods and technologies, especially computer vision, as Associate Professor Sergey Korchagin from the Department of Information Security of Automated Systems at the Institute of Electronics Engineering and Instrument Engineering at SSTU told Gazeta. ru.
To create a classifier, we used a database of images in which people express different emotions: joy, sadness, fear, anger, and others. The researchers then developed a neural network that can systematize a person’s emotions based on facial expressions.
The classifier can recognize 6 basic emotions: happiness (smile, bright eyes, positive mood), sadness (sad expression, sad eyes, and pessimistic mood), anger (alertness, pursed lips, tense facial expression), surprise (eyes wide open). , eyebrows raised, mouth wide open), fear (surprise, open eyes, pursed lips, facial tension) and disgust (grimace, gloomy look, rejecting expression).
The classifier can be used in medicine to diagnose mental disorders such as depression or anxiety. In marketing, emotion classifiers can be used to analyze and understand consumer responses to promotional materials, products or services. Emotion classifiers can also be used in the gaming industry. This will allow the creation of more realistic and engaging game scenarios or virtual worlds that respond to players’ emotional states.
Formerly Olga Kardymon, Research Fellow and Head of AIRI Bioinformatics Group said socialbites.ca, how Russian scientists taught the GENA neural network to analyze unlimited DNA sequences.