A Proposal Representation, Digital Coding and Clustering of Odor Information

In this paper, we propose techniques converting odor information into digital data to provide representation, digital coding, and clustering of various odors for olfactory information communication. About forty emotional adjectives describing various kinds of odors are selected as expressional receptors. Each of the odors is expressed by a definite adjective. The odors expressed using emotional receptors are coded using a predetermined method for transmission. After being transmitted, the odor data is decoded using a fuzzy c-means (FCM) clustering algorithm. Since the odor data are expressed in 40 dimensions, a longer computing times and a larger size of memory are required for FCM algorithm. To solve these problems, the dimensions of the data are reduced by principal component analysis (PCA) before using the FCM algorithm