Intelligent Learning for Computer Vision | 2021
An Optimal Feature Based Automatic Leaf Recognition Model Using Deep Neural Network
Abstract
Automatic recognition systems have recently gained enormous attentions form the research community. This is mainly because an automatic recognition system greatly reduces the human labor and intervention in the respective domains. It also eliminates the need of domain experts to a great extent. So, the success of an intelligent system nowadays inherently depends on the strength of its recognition system. In this paper, we have proposed a deep neural network based pattern recognition model for automatic classification of color leaves. The study initially uses eighteen features of the leaves for its basic model and thereafter has made a two step improvements on the initial model. The improvements have been achieved through dimension reductions and further through optimal feature selections. All of the variants of the proposed model have been tested with benchmark color leaf images and the results have been critically analyzed. The performances of the proposed model and its variants have been compared with that of six other models using standard indexes.