Yang Hu, Yukai Zhang, Shuai Liu, Guoxiong Zhou, Mingxuan Li, Yahui Hu, Johnny Li, Lixiang Sun
{"title":"DMFGAN:用于葡萄叶病识别的多特征数据增强方法。","authors":"Yang Hu, Yukai Zhang, Shuai Liu, Guoxiong Zhou, Mingxuan Li, Yahui Hu, Johnny Li, Lixiang Sun","doi":"10.1111/tpj.17042","DOIUrl":null,"url":null,"abstract":"<p><p>The use of deep learning techniques to identify grape leaf diseases relies on large, high-quality datasets. However, a large number of images occupy more computing resources and are prone to pattern collapse during training. In this paper, a depth-separable multifeature generative adversarial network (DMFGAN) was proposed to enhance grape leaf disease data. First, a multifeature extraction block (MFEB) based on the four-channel feature fusion strategy is designed to improve the quality of the generated image and avoid the problem of poor feature learning ability of the adversarial generation network caused by the single-channel feature extraction method. Second, a depth-based D-discriminator is designed to improve the discriminator capability and reduce the number of model parameters. Third, SeLU activation function was substituted for DCGAN activation function to overcome the problem that DCGAN activation function was not enough to fit grape leaf disease image data. Finally, an MFLoss function with a gradient penalty term is proposed to reduce the mode collapse during the training of generative adversarial networks. By comparing the visual indicators and evaluation indicators of the images generated by different models, and using the recognition network to verify the enhanced grape disease data, the results show that the method is effective in enhancing grape leaf disease data. Under the same experimental conditions, DMFGAN generates higher quality and more diverse images with fewer parameters than other generative adversarial networks. The mode breakdown times of generative adversarial networks in training process are reduced, which is more effective in practical application.</p>","PeriodicalId":233,"journal":{"name":"The Plant Journal","volume":null,"pages":null},"PeriodicalIF":6.2000,"publicationDate":"2024-10-24","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":"0","resultStr":"{\"title\":\"DMFGAN: a multifeature data augmentation method for grape leaf disease identification.\",\"authors\":\"Yang Hu, Yukai Zhang, Shuai Liu, Guoxiong Zhou, Mingxuan Li, Yahui Hu, Johnny Li, Lixiang Sun\",\"doi\":\"10.1111/tpj.17042\",\"DOIUrl\":null,\"url\":null,\"abstract\":\"<p><p>The use of deep learning techniques to identify grape leaf diseases relies on large, high-quality datasets. However, a large number of images occupy more computing resources and are prone to pattern collapse during training. In this paper, a depth-separable multifeature generative adversarial network (DMFGAN) was proposed to enhance grape leaf disease data. First, a multifeature extraction block (MFEB) based on the four-channel feature fusion strategy is designed to improve the quality of the generated image and avoid the problem of poor feature learning ability of the adversarial generation network caused by the single-channel feature extraction method. Second, a depth-based D-discriminator is designed to improve the discriminator capability and reduce the number of model parameters. Third, SeLU activation function was substituted for DCGAN activation function to overcome the problem that DCGAN activation function was not enough to fit grape leaf disease image data. Finally, an MFLoss function with a gradient penalty term is proposed to reduce the mode collapse during the training of generative adversarial networks. By comparing the visual indicators and evaluation indicators of the images generated by different models, and using the recognition network to verify the enhanced grape disease data, the results show that the method is effective in enhancing grape leaf disease data. Under the same experimental conditions, DMFGAN generates higher quality and more diverse images with fewer parameters than other generative adversarial networks. The mode breakdown times of generative adversarial networks in training process are reduced, which is more effective in practical application.</p>\",\"PeriodicalId\":233,\"journal\":{\"name\":\"The Plant Journal\",\"volume\":null,\"pages\":null},\"PeriodicalIF\":6.2000,\"publicationDate\":\"2024-10-24\",\"publicationTypes\":\"Journal Article\",\"fieldsOfStudy\":null,\"isOpenAccess\":false,\"openAccessPdf\":\"\",\"citationCount\":\"0\",\"resultStr\":null,\"platform\":\"Semanticscholar\",\"paperid\":null,\"PeriodicalName\":\"The Plant Journal\",\"FirstCategoryId\":\"2\",\"ListUrlMain\":\"https://doi.org/10.1111/tpj.17042\",\"RegionNum\":1,\"RegionCategory\":\"生物学\",\"ArticlePicture\":[],\"TitleCN\":null,\"AbstractTextCN\":null,\"PMCID\":null,\"EPubDate\":\"\",\"PubModel\":\"\",\"JCR\":\"Q1\",\"JCRName\":\"PLANT SCIENCES\",\"Score\":null,\"Total\":0}","platform":"Semanticscholar","paperid":null,"PeriodicalName":"The Plant Journal","FirstCategoryId":"2","ListUrlMain":"https://doi.org/10.1111/tpj.17042","RegionNum":1,"RegionCategory":"生物学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":null,"EPubDate":"","PubModel":"","JCR":"Q1","JCRName":"PLANT SCIENCES","Score":null,"Total":0}
引用次数: 0
摘要
使用深度学习技术识别葡萄叶片病害依赖于大量高质量的数据集。然而,大量图像会占用更多计算资源,而且在训练过程中容易出现模式崩溃。本文提出了一种深度分离多特征生成对抗网络(DMFGAN)来增强葡萄叶病数据。首先,设计了基于四通道特征融合策略的多特征提取块(MFEB),提高了生成图像的质量,避免了单通道特征提取方法导致的对抗生成网络特征学习能力差的问题。其次,设计了基于深度的 D-判别器,以提高判别能力并减少模型参数数量。第三,用 SeLU 激活函数替代 DCGAN 激活函数,以克服 DCGAN 激活函数不足以拟合葡萄叶病图像数据的问题。最后,提出了带有梯度惩罚项的 MFLoss 函数,以减少生成式对抗网络训练过程中的模式崩溃。通过比较不同模型生成的图像的视觉指标和评价指标,并利用识别网络对增强后的葡萄病害数据进行验证,结果表明该方法能有效增强葡萄叶片病害数据。在相同的实验条件下,与其他生成式对抗网络相比,DMFGAN以更少的参数生成了更高质量和更多样化的图像。生成式对抗网络在训练过程中的模式崩溃时间减少,在实际应用中更加有效。
DMFGAN: a multifeature data augmentation method for grape leaf disease identification.
The use of deep learning techniques to identify grape leaf diseases relies on large, high-quality datasets. However, a large number of images occupy more computing resources and are prone to pattern collapse during training. In this paper, a depth-separable multifeature generative adversarial network (DMFGAN) was proposed to enhance grape leaf disease data. First, a multifeature extraction block (MFEB) based on the four-channel feature fusion strategy is designed to improve the quality of the generated image and avoid the problem of poor feature learning ability of the adversarial generation network caused by the single-channel feature extraction method. Second, a depth-based D-discriminator is designed to improve the discriminator capability and reduce the number of model parameters. Third, SeLU activation function was substituted for DCGAN activation function to overcome the problem that DCGAN activation function was not enough to fit grape leaf disease image data. Finally, an MFLoss function with a gradient penalty term is proposed to reduce the mode collapse during the training of generative adversarial networks. By comparing the visual indicators and evaluation indicators of the images generated by different models, and using the recognition network to verify the enhanced grape disease data, the results show that the method is effective in enhancing grape leaf disease data. Under the same experimental conditions, DMFGAN generates higher quality and more diverse images with fewer parameters than other generative adversarial networks. The mode breakdown times of generative adversarial networks in training process are reduced, which is more effective in practical application.
期刊介绍:
Publishing the best original research papers in all key areas of modern plant biology from the world"s leading laboratories, The Plant Journal provides a dynamic forum for this ever growing international research community.
Plant science research is now at the forefront of research in the biological sciences, with breakthroughs in our understanding of fundamental processes in plants matching those in other organisms. The impact of molecular genetics and the availability of model and crop species can be seen in all aspects of plant biology. For publication in The Plant Journal the research must provide a highly significant new contribution to our understanding of plants and be of general interest to the plant science community.