Scientists are Working on Determining Emotion from Human Brain Signals at GTU
02-09-2024
A group of scientists from the Georgian Technical University and a graduate of the Automation and Test Engineering master’s program of the Faculty of Informatics and Control Systems, a senior teacher of the faculty, and a researcher of the Institute of Computational Mathematics named after Niko Muskhelishvili of GTU - Natia Kukhilava is working on determining emotion from human brain signals.
The researcher of the Georgian Technical University is one of the members of the Forbes 30 Under 30 Europe 2024 list.
The publication named her among the winners in the category of technology and e-commerce. Together with German and French colleagues, Natia Kukhilava is working on determining emotion from human brain signals (EEG) using deep learning.
“What does it mean to recognize emotion from human brain signals? The human brain is constantly active and emits signals based on what a person is doing or feeling. An electroencephalogram (EEG) is a device that can read these signals, which are processed by scientists.
In several leading research institutes of the world, people’s brain signals were recorded according to the following principle: a person watches a video and then evaluates himself/herself and what emotion he/she experienced while watching this video. During the experiments, several people were recorded, who watched videos.
As a result, we have a dataset with EEG signals, and from each signal, we know what emotion a person experienced at that time. The goal of the research is to create a better AI model than all existing techniques that will be able to predict (guess) what emotion a person is experiencing at a particular moment from the EEG signal. Today, we have already created a framework where all datasets and AI models known in this field are embedded, it is possible to add any new model or dataset here.
As a rule, building models from articles and finding them in datasets requires a lot of time and resources, so when in the research process you want to compare your results with the results of other articles/models/datasets, their implementation is too difficult and takes much time. The framework we created allows scientists to conduct experiments using different AI models and datasets much more easily, compare the results with each other, and speed up the research process.
The framework has already been implemented in the teaching process of Saarland University in Germany, where master’s students use it to conduct their research. In addition, we have been invited to the ACII 2024 conference in Glasgow, where we organized a workshop, the main topic of which is emotion recognition and this framework,” said Natia Kukhilava.
For information: the startup of the researcher Natia Kukhilava of the Georgian Technical University - Helio.AI - became the GITA grant winner. The researcher is also involved in the Horizon Europe project GAIN, where she works with German and French colleagues and GTU scientists from human brain signals (EEG) using deep learning to determine emotion. Within this project, project participants are invited to the ACII 2024 conference as workshop organizers.
The researcher of the Georgian Technical University is one of the members of the Forbes 30 Under 30 Europe 2024 list.
The publication named her among the winners in the category of technology and e-commerce. Together with German and French colleagues, Natia Kukhilava is working on determining emotion from human brain signals (EEG) using deep learning.
“What does it mean to recognize emotion from human brain signals? The human brain is constantly active and emits signals based on what a person is doing or feeling. An electroencephalogram (EEG) is a device that can read these signals, which are processed by scientists.
In several leading research institutes of the world, people’s brain signals were recorded according to the following principle: a person watches a video and then evaluates himself/herself and what emotion he/she experienced while watching this video. During the experiments, several people were recorded, who watched videos.
As a result, we have a dataset with EEG signals, and from each signal, we know what emotion a person experienced at that time. The goal of the research is to create a better AI model than all existing techniques that will be able to predict (guess) what emotion a person is experiencing at a particular moment from the EEG signal. Today, we have already created a framework where all datasets and AI models known in this field are embedded, it is possible to add any new model or dataset here.
As a rule, building models from articles and finding them in datasets requires a lot of time and resources, so when in the research process you want to compare your results with the results of other articles/models/datasets, their implementation is too difficult and takes much time. The framework we created allows scientists to conduct experiments using different AI models and datasets much more easily, compare the results with each other, and speed up the research process.
The framework has already been implemented in the teaching process of Saarland University in Germany, where master’s students use it to conduct their research. In addition, we have been invited to the ACII 2024 conference in Glasgow, where we organized a workshop, the main topic of which is emotion recognition and this framework,” said Natia Kukhilava.
For information: the startup of the researcher Natia Kukhilava of the Georgian Technical University - Helio.AI - became the GITA grant winner. The researcher is also involved in the Horizon Europe project GAIN, where she works with German and French colleagues and GTU scientists from human brain signals (EEG) using deep learning to determine emotion. Within this project, project participants are invited to the ACII 2024 conference as workshop organizers.
