Impact of Generative AI in Diagnosing Diseases in Agriculture
Yuvpartap Singh Klair, Kushagra Agrawal, Ayush Kumar
Abstract
In the domain of agriculture, accurate disease diagnosis is crucial for ensuring crop health and maximizing yield. However, the scarcity of diverse and well-annotated datasets poses challenges to developing robust diagnostic models. This research explores the potential of Generative AI to address these limitations and enhance the accuracy of disease diagnosis. To address the issue, this research offers a novel method of data augmentation that applies a cutting-edge generative AI algorithm, namely Generative Adversarial Networks (GANs), to a subset of the “Paddy Doctor” dataset, which consists of 10,407 photos of paddy crops suffering from various diseases. GANs are employed to generate synthetic images that seamlessly integrate with the original dataset, effectively enhancing its size and diversity. Leveraging the adversarial training process, this method produces realistic images capturing detailed features of diseased crops. Numerous studies demonstrate the effectiveness of this method in enhancing the applicability and accuracy of crop disease classification models. This method proves valuable in addressing data scarcity issues within the agricultural domain.