Olfactory information technology is recently evolving. It is intended to handle even olfaction in cyberspace. First, odor reproduction technique is presented. Odor can be approximated using small number of odor components when we adjust the mixture composition. The technique to extract odor components and the experimental data will be presented, Then, the technique of predicting odor impression is presented. Odor impression made up of several odor descriptors is predicted from mass spectrum using deep learning. Natural language processing is also used to characterize odor descriptors. Finally, an olfactory display to blend multiple ingredients is explained. We recently evaluated the reproduced smell from olfactory display using sensory test. The principle of olfactory display and result of sensory test are presented in this conference.
各種検索
サポート
ヘルプ