Odor Synthesis Method Based on Odor Components Using Deep Neural Networks

カテゴリ: 研究会(論文単位)

論文No: CHS24004

グループ名: 【E】センサ・マイクロマシン部門 ケミカルセンサ研究会

発行日: 2024/02/27

タイトル(英語): Odor Synthesis Method Based on Odor Components Using Deep Neural Networks

著者名: Aleixandre Manuel(科学技術創成研究院),Prasetyawan Dani(科学技術創成研究院),中本 高道(科学技術創成研究院)

著者名(英語): Manuel Aleixandre(Institute of Innovative Research),Dani Prasetyawan(Institute of Innovative Research),Takamichi Nakamoto(Institute of Innovative Research)

キーワード: Odors|Deep neural network|Gradient descent|Odor synthesis|Odor components|Mass spectroscopy

要約(日本語): A method for odor analysis and synthesis is studied. A DNN to predict odor descriptors using mass spectrometry as sensor data was trained first. Then it was used to create various odor recipes by blending essential oils using gradient descent. The approach was validated by several methods including sensory tests.

要約(英語): A method for odor analysis and synthesis is studied. A DNN to predict odor descriptors using mass spectrometry as sensor data was trained first. Then it was used to create various odor recipes by blending essential oils using gradient descent. The approach was validated by several methods including sensory tests.

本誌: 2024年3月1日ケミカルセンサ研究会

本誌掲載ページ: 15-20 p

原稿種別: 英語

PDFファイルサイズ: 1,622 Kバイト