International Journal of Contemporary Research in Multidisciplinary

The rapid advancement of artificial intelligence has significantly transformed the domain of plant disease detection and classification, offering precise and efficient diagnostic solutions. This survey critically examines recent developments in transfer learning and generative models for leaf disease detection. Transfer learning methods, particularly those leveraging deep convolutional neural networks (CNNs) and pre-trained architectures, have demonstrated exceptional classification performance, with accuracies reaching up to 96.25%. Similarly, generative models such as Generative Adversarial Networks (GANs) and Variational Autoencoders (VAEs) have been successfully employed for dataset augmentation and feature learning, enhancing the robustness of classification models. By synthesizing findings across multiple studies, this work highlights the diversity of methodologies, identifies the most effective techniques, and discusses their implications for precision agriculture. Furthermore, the paper provides a comparative evaluation of both paradigms, offering valuable insights into their potential convergence for scalable and sustainable plant disease diagnostics.