Development of Kansei Fuzzy CBIR System
Shigeo Wada, Zhuo Liu · Medical Entomology and Zoology · 2003
In this report, a Kansei fuzzy CBIR (Content Based Image Retrieval) system is developed. A novel CBIR model based on Kansei engineering is proposed. The image features are extracted using wavelet packet transform. The feature vectors are reduced to low dimensional feature vector using FDA (Fuzzy Discriminate Analysis). The desired image reflecting user’s impression and sensitivity can be retrieved flexibly by using fuzzy reasoning distance. In the simulations, color texture image databases are constructed. The effectiveness of the proposed system is demonstrated with the databases.