基于生成对抗网络的数据增强技术在葡萄叶数据增强中的应用

Application of Generative Adversarial Network Based Data Enhancement Techniques in Data Enhancement of Grape Leaf

研究生成对抗网络(Generative Adversarial Networks,GAN)在葡萄叶数据增强中的应用。增强的葡萄叶数据可以提升病害检测的准确性。传统的数据增强方法虽然能在一定程度上缓解数据不足的问题,但是生成的图像往往缺乏多样性和真实感。为此,结合GAN和多尺度注意力机制设计一种改进的对抗生成网络AMSGAN。通过训练该模型的生成器和判别器,最终可生成高质量、多样性的葡萄叶图像。实验结果验证了该模型在葡萄叶数据增强中的有效性和潜力。

This paper studies the application of Generative Adversarial Networks(GAN)in the enhancement of grape leaf data.Enhanced grape leaf data can improve the accuracy of disease detection.Although traditional data enhancement methods can alleviate the problem of insufficient data to a certain extent,the generated images often lack diversity and realism.Therefore,in this paper,an improved adversarial generation network AMSGAN is designed by combining GAN and multi-scale attention mechanism,By training the generator and discriminator of this model,high-quality and diverse grape leaf images can be generated finally.The experimental results demonstrate the effectiveness and potential of this model in enhancing grape leaf data.