Asian Journal of Computer and Information Systems
Faculty of Health, Engineering and Science
School of Computer and Security Science/eAgriculture Research Group
This article details the effect of Gaussian smoothing parameter (spread) on the performance of Probabilistic Neural Networks (PNN). Two (2) different Genetic Algorithms (GAs) were used to optimize the PNN spread in order to avoid under and over fitting. In this work there is a novel combination of Cellular Neural Networks (CNN), Probabilistic Neural Networks (PNN) and GA to address the present challenges on automatic identification of plant species. Such problems include misclassification species of plants that are similar in shapes and image segmentation speed. In this work, GA was used in both feature selection and PNN parameter optimization. The GA developed herein improved the performance of the PNN. This work serves as a framework for building image classification or pattern recognition system.
Creative Commons License
This work is licensed under a Creative Commons Attribution 3.0 License.