Color Inference Using an Enhanced Fuzzy Method

Abstract

Color information recognition methods based on the RGB color model, which is designed on the basis of static fuzzy inference rules, are being widely used at present. However, these methods have certain limitations because of the nature of the model used: detachment of human vision and limited choice of environment. In this paper, we propose a method based on the HSI model and a new inference process that resembles the human vision recognition process. This method allows the user to add, delete, or update inference rules. In our method, membership intervals are designed with sine and cosine functions in the H channel and trigonometric style functions in the S and I channels. The membership degree is computed via an interval merging process. Then, inference rules are applied to the result in order to infer the color information. Experimental results show that our method is more intuitive and efficient than that based on the RGB model.

Interaction between humans, NAO robot and multiple cameras for colored objects recognition using information fusion

“…The related results are not detailed in this paper. For each demanded color Ci and for each camera, we have the combined mass defined by the following: (8) The parameter k represents the conflict between the s sources, and Yj is the information given by the source j.…”

“…Indeed, this can describe perfectly well the link between input image and color label in terms of human reasoning, the latter is more widely used. In [6][7] [8], Fuzzy Logic was used to classify the pixels into color labels. Firstly, they defined the set of membership functions for each dimension of a color space; those functions are used to fuzzify the input pixels into linguistic terms.…”

Color Recognition for NAO Robot Using Sugeno Fuzzy System and Evidence

“…Since the color recognition is executed under uncertain and imprecise environments where an exact analytical procedure design cannot be determined, artificial intelligent (AI) methods like fuzzy systems and/or neural networks [10] [12] seem to be very effective tools to develop color recognition strategies. Recently, such intelligent techniques have been successfully used for color recognition in RGB, HSV and Lab spaces.…”