Detecting plant leaf diseases through a Constitutive Artificial Neural Network that utilizes Gabor filters

The appearance of novel diseases affecting plant foliage represents a significant threat to agricultural productivity and global food stability. Swift and precise identification of these plant ailments is imperative, as it mitigates unnecessary costs and minimizes economic setbacks and environmental impact. This research introduces a technique named Constitutive Artificial Neural Network (CANN) for Detection of Plant Leaf Disease in conjunction with Gabor filters. Leveraging Gabor filter feature extraction improves. The precision of diseased plant leaves identification. The proposed approach integrates Gabor filters and Constitutive Artificial Neural Network within an ensemble classification framework to enhance classification results. Research has demonstrated that integrating Gabor filters utilizing Constitutive Artificial Neural Network leads to enhanced accuracy in disease detection. This approach utilizes Gabor filters to extract textural details from leaf images, facilitating the differentiation between healthy and diseased leaves. Overall, combining Constitutive Artificial Neural Network with Gabor filter feature extraction boosts the precision of plant leaf disease identification. The process involves enhancing images with a Gabor Filter and eliminating noise. Subsequently, the Adaptive Convex Clustering (ACC) method segments the images, after which Fast Fourier along with Continuous Wavelet transformation are used for feature extraction. Lastly, Constitutive Artificial Neural Network (CANN) categorizes the input image into various groups, comprising healthy leaves and diverse disease types like Powdery mildew, White Blight, and Downy mildew and healthy pumpkin leaf images and achieved 99.48% significant accuracy.