Classification of Rice Using Genetic Fuzzy Cascading System

Classification can be done using various AI methods currently available in the literature. However most of the AI techniques are black boxes. We do not know what is going on inside them and hence explainability of the model is very limited. A fuzzy system can increase the explainability to a certain degree. In this paper, we classify two different types of rice with the use of a genetic fuzzy cascading system and compare the accuracy with other methods. A total of rice grain images are converted to greyscale images and with help of computer vision, the attributes are obtained. These data are used for classification using a 7 input 2 output Fuzzy Inference System (FIS) with multiple levels of cascading. Each of the input, output membership functions and the rule base are tuned using Genetic Algorithm. Our current approach was able to produce 94% accuracy in the validation set.