Support System for Multimedia Information Data Acquisition Based on Fuzzy Inference with a Fuzzy Shift
Kabsuk Oh*, Kaoru Hirota**
*Department of Computer Science, TongMyong College,
505 Yongdang-Dong, Nam-Gu, Pusan 608-737, Korea.
**Interdisciplinary Graduate School of Science and Engineering, Tokyo Institute of Technology, 4259 Nagatsuta, Midori-ku, Yokohama 226-8502, Japan.
Support system construction for multimedia information data acquisition based on fuzzy inference with a concept of fuzzy shift is proposed, where the multimedia means the five senses. Observed information from the outside world is characterized by VAGOT (visual, acoustic, gustatory, olfactory, and tactile) time series data. Here, multimedia information centers on image and sound are represented by membership functions. Fuzzy rules based on visual and acoustic information are used to identify the appropriate time interval on multimedia input data. The proposed system is constructed by rule construction, data acquisition, a rule base, and a support system. When an instruction of the user has temporal vagueness, rule base is updated by the support system based on fuzzy shift. The method is applied to an experiment related to the decision of a temporal boundary subset of vehicles and results verify the feasibility of the proposed method.