Automated recognition of Ficus deltoidea using ant colony optimization technique

Improving in the fields of soft computing and artificial intelligence, the branch study of automated herb recognition among plenty of weeds has become challenging issue due to their applications in medicine, food and industry. This paper presents innovative method to improve the accuracy of classification as well the efficiency, such that irrelevant features that make computational complexity are ignored by feature subset selection that is proposed by means of ant colony optimization algorithm (ACO). At first, through image processing specified features are extracted from the Ficus deltoidea leaves such as vein, morphology and texture features and they construct a search space to be chosen for the optimal subset features that is selected by ACO algorithm as support vector machine (SVM) classify them. The experimental results have shown that the proposed method not only optimize feature subset, but also has a remarkable positive impact on accuracy.