“A flavor sensor, as its name suggests, measures flavors. All flavors are thought to be quantifiable using five basic ones – sweet, salty, sharp, bitter, and savory. So what we do is evaluate these five flavors using five gradations.”
“Our system is especially original in the way it measures “sweet” and “savory.” Other companies offer flavor sensors, but ours uses the amperometric method, where an electric current is measured when sweet and savory flavors are checked. Because this method directly measures the substances that produce sweet and savory flavors, its advantage is that it provides higher precision.”
Additionally, this system uses neural network analysis, which simulates biological neurons, to quantify flavors. This achieves a sensor that more closely resembles the human sense of taste.
“For example, if you add sugar to bitter coffee, the quantity of the bitter constituent doesn’t change, but our tongues sense that the coffee is less bitter. That’s precisely because we have non-linear nerve circuits. So we thought that with a neural network, like those used in control engineering, we could make the sensor respond more like a person. Using the neural network, our flavor sensor properly reproduces the perception that the coffee becomes less bitter and more sweet.”
This flavor sensor technology is currently utilized widely, not just in research, but in marketing and other fields.
“This sensor provides solutions; for example, “What snacks would taste good with Awamori?” We think that the value lies in how to utilize the data, rather than just the data itself. In fact, we’ve had a variety of business inquiries, and the range of applications is also growing.”