A Novel Fuzzy Expert System Using Image Processing For Sale Car Shape With Online Membership Function

Abstract

In This paper we proposed a novel fuzzy retrieval system for purchasing cars employing image processing. The car shape is an important factor when selecting a car type. This system aims to support persons who are not good with cars. When we try to purchase a car, they can use this system easily as if they ask casually someone else who knows more about car. Unspecific conditions are expressed by the fuzzy set, and the level matching conditions are expressed by the grade values. To use this more practically, a GUI form with selection menus is developed. This system is designed to change membership function automatically for improving usability of this system. In addition, calculating curvature by the car shape using the image processing, and adding items for selecting a car shape from roundness and sharpness. Furthermore, the effectiveness of applying fuzzy logic to express man's subjectivity when selecting car type.