Comparative analysis of Visual Recognition Capabilities of CNN Architecture Enhanced with Gabor Filter

Convolutional Neural Networks have long been regarded as potent tools for computer vision and deep learning. They have proven to be effective in analysis of visual imagery with increasing accuracy. Another tool which has profoundly impacted the performance of CNN has been the use of Gabor Filters. Herein, comparative analysis survey is performed on the performance of CNN architecture, with and without application of Gabor Layers, on datasets of plant diseases to identify potential in agricultural applications of these mechanisms. The aforementioned dataset consists of 38 flora classes with an additional class to facilitate identification of absence of plants. Hence, it not only conduct a survey of processes concerned with image processing and disease identification but also probe the established theories in plant disease identification through machine learning concepts.