Trial for Quantifing Kansei by Using Image Processing

Abstract

In this study, we will try to quantify kansei by using image processing. First, we will make image of rgb level of 2 dimension into 1 dimension and assume it as time series data. Then we will go through signal processing like FFT and derive 1/f deviation or dimension of fractal to quantify the image. Then, we will compare these values with kansei key words and derive some relationship between them to design a new product. In this study, we use this system to inside door for apartment house.