A heuristic model for optimizing fuzzy knowledge base in a pattern recognition system
Amir Jamshidnezhad, J. M. Nordin · Journal of Scientific & Industrial Research · 2012
This study presents a genetic algorithm (GA) to optimize performance of a fuzzy system for reconition of facial expressionfrom images. In proposed model, a Mamdani-type fuzzy rule based system recognizes emotions, and a GA is used to improveaccuracy and robustness of the system. To evaluate system performance, images from FG-Net (FEED) and Cohn-Kanadedatabase were used to obtain the best functions parameters. Proposed model under training process not only increased accuracyrate of emotion recognition but also increased validity of the model in adverse conditions.