Yasue Mitsukura

Professor | Keio University

Transcript of the presentation Real Time Emotion Detection Using EEG With Real Time Noise Reduction, given at the NTT Upgrade 2020 Research Summit, October 1, 2020

Hello, nice to meet you! My name is Yasue Mitsukura, I’m a professor in Keio University in Japan. So, today I want to introduce my research that title is: A Real Time Emotion Detection using EEG with real time noise reduction. First of all, I want to introduce myself, My major is System Identification and Signal Processing on noise removed in the biomedical signal process. For all of them, a common technique, is mathematical modeling by equation using the self-tuning identification method. So, today’s topic is EEG model by the various Concept for his heavy noise. We call this technique the KANSEI Model.


Now what is a KANSEI? KANSEI is Japanese verb, because studied at first in Japan. So KANSEI is similar to emotion and sensibility, but they are quite different. Thus emotion and sensibility is innate ability, but KANSEI is acquired after birth. KANSEI is similar to how to feel. So we focus on this KANSEI using a brain signal.


As for the brain signal, there are many ways to know the brain; for example, the optical reading X-CT, MRI, MEG, EEG, Optical Topography and IMRI. By using these devices, we have three areas of research, for example: the Neuro Engineering area, for application, coding and (indecipherable); Neuro Science area, for understanding the mechanism; and a medical reward area for treatment. So, but it’s very important to use depending on the purpose.


So what data can be obtained? In the case of EEG we can see the activity of neuron in the scalp. In the case of NIRS, we can obtain the level of oxygen in the blood brought. In the case of Nitro and Encephalogram, we can see the activity of neuron by contactless. In the case of Positon-Emission Topography, we can get activity of reception by the contactless. If we use fMRI, we can measure the amount of blood by the contactless. These devices are showing in these figures. So, our motivation is to get the KANSEI equation, using the model by System Identification with noise removal. And the second motivation is to realize this simple and smart concept selection using the EEG information.


When we use the fMRI, the large scale and the expensive and binding, it is un-useful. So we focus on the EEG because the EEG in small scale, inexpensive, and nonbinding and useful. So we focus on EEG. So EEG is action potential measured from the scalp. Detected data is translated to the frequency domain. In the frequency domain, that 0.4 to 4 we call the delta wave; 4 to 6, we call that theta wave; and 7 to 14 are called the alpha wave; and 14 to 26 we call that beta wave. In combination method, we want, if we want to use the [unintelligible] as deep sleep, we use wave delta wave; in a case of light sleep, we use a theta wave, and so on. But this is just a new assessment method. So, we cannot use that for all the film actually the accuracy is under the 20%. So we need to define the situation originally.


So call this technique KANSEI Model. So these are the block diagram. You can see KANSEI model. So, this filter, this part is for the noise; this part for the mathematical model. So, we calculate this transfer function like this this is a discrete time model and this term is continuous time model.


So, then we rewrite this part to discrete time model. So, we calculate this part, like this; this first part and the second part is calculated by the Pade approximation. So, we can get this the Augmented model. So, then that we rewrite this part by using that Z transfer function transformation. So, we will write these augmented models like this.


So, the alpha of the inverse and beta of the inverse is defined as this equation. So, each coefficient is calculated by this equation, and then we calculate a1 a2, ba  b2 by using this because the least squares algorithm. So we call this identification method Self Tuning Identification method.


This is an example for stress modeling. First of all, we decide, we gather that data deck stress model. So, we move the small beans tray to tray for one hour. So, last 10 minutes, we used as stress and we measure that [culture?] of stress level, and we associate the EEG and we measure the 8,000 data in 17 years, we get at 17 years. So, in the case of the simple easy devices, there are many simple devices in the world like this. So, many over them we calculate the signal to noise ratio, the signal means their medical EEG system, and the each device measure is in ratio. And then we investigate 58 kinds of devices, and almost of all devices are noise devices.


So, I’m often asked about to various persons what device is the best? So, my answer is anything our skill is signal processing and if we’ve got a lot of data can be obtained from the device, no matter what device we may use, the same result can be shown. Our analogy: is the raw signal processing and our system is structured by 17 years data par one situation. So, my answer is what, anything. So, we apply the system to a real product; we call this product KANSEI Analyzer.


In KANSEI Analyser you can see the KANSEI that is right there. They are called interests, sleepiness, concentration and the like. So that we combine that this KANSEI Analyser and that camera system, we make the Neuro system NEUROCAM. so please show it (demonstrations) this is the neuro cam.


This is the EEG system and there we can get KANSEI by using that iPhone and we combine the camera system by the iPhone camera and if the concept is higher than 60%, So, automatically record it like this. (demonstrates) So, every time that we wear the EEG devices, we can know the awareness, KANSEI model. (demonstrates) So, finally, we combine the each of KANSEI, like this movie. So, we can see the one day KANSEI the movie. So, this is NEUROCAM.


And the next example is Neuro-marketing using a KANSEI analyzer. So, this is No.1 CM Japan. (a movie plays in Japanese.) But we don’t know what is the number one point. So then we analyze the deeds CM by using the KANSEI analyzer. So we can get the real time concept. Then we can see the one-by-one situation like this. So this is the interest level, and we can see the high interest like this. So the recorded moment automatically.


And the next one is Real Application (product design), News Reporter: Japanese professor, has come up with a new technology she claims can read minds. She says the brain wave analysis system will help businesses better understand their customer’s needs.

Video plays in the news room: Workers at a major restaurant chain are testing a menu item that is being developed. This device measures brainwaves, from the frontal lobes of people who try the product An application analyzes five feelings, how much they like something and their interest, concentration, stress, and sleepiness. The new menu item is a cheese souffle, topped with Kiwi orange and other fruit. The App checks the reaction of a person who sees the souffle for the first time.


>> Please open your eyes.

>> News Presenter: When she sees the souffle, the likes and interests feelings surge on the ground. This proves the desert is visually appealing.

>> Now, please try it.

>> News Presenter: After the first bite, the like level goes up to 60, that shows she likes how the dessert tastes. After another bite, the like level reaches 80 she really enjoys, the taste of the souffle. It scores high in terms of both looks and taste. But there’s an unexpected problem. When she tries to scoop up the fruit, the stress level soars to 90.

>> I didn’t know where to put the spoon, I felt it was a little difficult to eat. It turned out it was difficult to scoop up the fruit with the small spoon. So people at the restaurant chain, are thinking of serving this a plate with a fork instead.

>> What’s the big difference? With the device we can measure emotional changes in my new detail in real time.

>> News Presenter: This is a Printing and Design firm in Tokyo It designs direct mail and credit card application forms. The company is using the Brain Wave Analyzing System to improve the layout of its products. The idea is to make them easier to read. During this test, the subject wears an eye tracking device to record where she’s looking, in addition to the Brain wave Analyzing device. Our eye movements are shown by the red dots on this screen. Stress levels are indicated on the graph on the left.

>> Please fill out the form.

>> News Presenter: This is a credit card application form. Right after she turns her eyes to this section, her stress levels shoots up. It was difficult to read, as each line contains 60 characters. So, they decided to divide the session in two, cutting the length of the lines by half.

>> This system is very useful for us; we can offer a differentiated service to our clients by providing science-based solution

>> News Presenter] The brainwave Analyzer …

>> Now, we can select the KANSEI detection like this, like or dislike concentration, interest, sleepiness, stress, continue, like, comfortable, uncomfortable unpleasant, relax, devotion degree, addiction, taste, riding comfort, satisfaction and achievement.


So, finally, by concluding my presentation. So, in this presentation, we introduce our research, we construct the KANSEI equation, and we demonstrate strict signal processing, and we apply the proposed method to a real product, we named the KANSEI analyzer.


So this is the first in the world. That’s all. Thank you so much.

Real Time Emotion Detection Using EEG With Real Time Noise Reduction

Yasue Mitsukura head shot

Yasue Mitsukura
Professor | Keio University