ITMEDIA NEWS with 95 % accuracy (ITMEDIA NEWS) --Yahoo! News that sniffs whiskey.

Appeared to use this method to use

　中国のShandong First Medical University、オーストラリアのUniversity of Technology Sydney、中国のXidian University、カナダのUniversite du Quebec a Trois-Rivieresによる研究チームが開発した「The Use of Electronic Nose for the Classification of Blended and Single Malt ScotchWhisky Is a category that identifies the differences between brands, locations, and styles (such as blended, malt, etc.) simply by smelling the Scotch whiskey. [Image] The outline of the system is identified with an average accuracy of about 95 %. The proposal method is that it does not require complicated sample preparation and can conduct quickly and test on site. In some cases, malicious people change the contents and notation of whiskey, or do fraud or counterfeit, such as mixing impurities. Training experts and experienced enthusiasts may be able to distinguish the difference in whiskey from the scent, but it will be difficult for most people to distinguish only with scents. Until now, devices and methods that identify from the smell have been proposed, but the hurdles are generally used because the device is expensive, the sample preparation and usage are complicated. The classifier proposed this time is cheaper and easier to use than before, but simply smells whiskey to identify brands, locations, and styles in a short time. The setup is the classification prototype "NOS.E "(a system based on five standard MOS gas sensors), a notebook PC that connects to it, and a small bottle to put whiskey. The scent of whiskey in a small bottle is injected into the gas sensor, and the sensor senses various smells and sends data to the computer and analyzes it. Then, the characteristics of the important scent are extracted, and the classification is performed with the machine learning algorithm designed to recognize the brands, locations, and styles of whiskey. In the experiment, from JohnnyWalker's red label / black label to 12 years of McCaran, three types of blended malt and three types of single malt whiskeys are used, and these devices are used for brands, production areas and styles.We conducted a test to identify based on the basis. Compared with the results obtained by GC x GC-TOFMS (comprehensive 2D gas chromatography, flight time type mass analysis) as a confirmation analysis. The GC device is highly accurate, but the usage is complicated and expensive. As a result, this method does not require an advanced procedure, but is 96.15 %, 100 % produced, style 92.It turned out that I could identify with 31 % high accuracy.Furthermore, as a sample, a demonstration of the field test was performed using 58 whiskey datasets collected at the CEBIT 2019 exhibition, and the company, the production area, style, and high classification accuracy were achieved, emphasizing its effectiveness.Source and Image Credits: w.Zhang, T.LIU, A.Brown, m.UELAND, S.L.Forbes and S.W. Su, “The Use of Electronic Nose for the Classification of Blended and Single Malt ScotchWhisky,” in IEEE Sensors Journal, doi: 10.1109/JSEN.2022.3147185.* Mr. Yuki Yamashita, who presides over the web media "SEAMLESS", introducing the latest research on technology.Mr. Yamashita picks up and explains a highly new nature paper.

ITMEDIA NEWS