Scholarlink

Timely detection of wheat rust diseases emerges as a critical step for safeguarding wheat crops, which serve as a major source of food for the majority of population worldwide. An automated system is required for early prediction of wheat disease. This paper introduces a deep learning based model, ViT-GAN, a fusion of Vision Transformers and Generative Adversarial Networks, adopting few-shot and transfer learning techniques. ViT-GAN, optimized for wheat rust detection, utilizes the power of generative models to augment datasets with synthetic wheat leaf images, while the few-shot capabilities enable accurate disease detection. Further, ViT-GAN’s optimization and integration of transfer and few-shot learning techniques facilitate the way for adaptable and efficient AI solutions in diverse domains with limited labeled data.

错误：搜索内容不能为空，请输入英文关键词
错误：关键词超出字数限制，请精简
高级检索

Optimizing Vision Transformer Based Generative Adversarial Networks (ViT-GAN) for Early Detection of Wheat Rust Disease Utilizing Few-Shot and Transfer Learning

  • 会议论文

  • 发表日期：

    2025年

查看全文
  • K. Vishnu Vardhan,
  • H. Ateeq Ahmed

摘要

Timely detection of wheat rust diseases emerges as a critical step for safeguarding wheat crops, which serve as a major source of food for the majority of population worldwide. An automated system is required for early prediction of wheat disease. This paper introduces a deep learning based model, ViT-GAN, a fusion of Vision Transformers and Generative Adversarial Networks, adopting few-shot and transfer learning techniques. ViT-GAN, optimized for wheat rust detection, utilizes the power of generative models to augment datasets with synthetic wheat leaf images, while the few-shot capabilities enable accurate disease detection. Further, ViT-GAN’s optimization and integration of transfer and few-shot learning techniques facilitate the way for adaptable and efficient AI solutions in diverse domains with limited labeled data.