|
Abstract
|
Deep learning models have demonstrated remarkable success in addressing a wide range of real-world problems. Specifically, “convolutional neural networks (CNNs)” have proven to be effective in various computer vision tasks. However, determining the “optimal architecture” and “influential hyper-parameters” of a CNN is a challenging task due to the complex and extensive search space involved. To overcome these challenges, we propose utilizing the antlion optimization (ALO) algorithm to design a CNN. This approach enables us to obtain “optimal architectures” and identify the most effective “hyper-parameters” required for training the CNN. By the recent developments made in the agriculture industry, particularly, rapid ongoing variation in the “foundation” as well as “agriculture patterns,” novel illnesses are continually emerging on the plants’ leaves. Accordingly, farming and food production will be faced with real dangers all around the globe. To address this important threat, we applied our proposed deep learning-based approach. To this end, we developed new CNNs to recognize different diseases associated with various plant species. Extensive experiments on four challenging plant leaf disease datasets were conducted to judge the functionality of our proposal. Our acquired performances convey the domination of our proposal against state-of-the-art approaches.
|