针对于当前人脸遮挡修复方法中出现修复图像信息不完整、纹理模糊、产生伪影、细节欠佳以及模型训练不稳定等问题,提出一种基于CSWin-Transformer和WGAN的人脸遮挡修复方法。该方法以Encoder-Decoder结构作为生成器,在生成器中引入CSWin...针对于当前人脸遮挡修复方法中出现修复图像信息不完整、纹理模糊、产生伪影、细节欠佳以及模型训练不稳定等问题,提出一种基于CSWin-Transformer和WGAN的人脸遮挡修复方法。该方法以Encoder-Decoder结构作为生成器,在生成器中引入CSWin-Transformer Block来精细识别和处理被遮挡的面部区域,以提高处理的针对性和效率,解码器通过跳跃连接与编码器多尺度特征融合,更好学习图像的细节特征,优化最终效果。在判别器中引入Wasserstein距离,来提高模型训练稳定性以及生成图像的真实性,同时在判别器中引入CSWinSelf-Attention,增强判别器对图像全局结构和细节信息的理解。实验结果显示,文章方法在所使用的CelebA的数据集上有良好的修复效果,在峰值信噪比(PSNR)和结构相似性指数(SSIM)指标上与目前一些图像修复方法相比表现更优。 In view of the problems of incomplete repair image information, blurred texture, artifacts, poor details and unstable model training, a face occlusion repair method based on CSWin-Transformer and WGAN is proposed. This method takes Encoder-Decoder, structure as the generator, and introduces CSWin-Transformer Block in the generator to finely identify and process the occluded face areas, so as to improve the pertinacity and efficiency of processing. The decoder integrates with the encoder multi-scale features through jump connection to better learn the detailed features of the image and optimize the final effect. The Wasserstein distance is introduced into the discriminator to improve the stability of the model training and the authenticity of the generated image. Meanwhile, CSWin Self-Attention is introduced in the discriminator to enhance the understanding of the global structure and details of the image. The experimental results show that the method has good repair effect on the data set of CelebA used, and better than some current image repair methods in peak signal-to-noise ratio (PSNR) and structural similarity index (SSIM) indicators.展开更多
针对水稻病害图像数据集样本较少而影响深度神经网络模型学习的精度问题,提出一种改进的对抗生成网络模型ViT-WGAN-GP(Vision Transformer and Wasserstein Generative Adversarial Networks with Gradient Penalty)用于对图像数据集进...针对水稻病害图像数据集样本较少而影响深度神经网络模型学习的精度问题,提出一种改进的对抗生成网络模型ViT-WGAN-GP(Vision Transformer and Wasserstein Generative Adversarial Networks with Gradient Penalty)用于对图像数据集进行增强。首先在生成模型引入Vision Transformer结构加强对全局特征的学习;其次在判别模型采用WGAN-GP结构,通过Wasserstein衡量函数和梯度惩罚项保证模型训练的稳定性,提升生成图像的效果;最后使用增强后的样本集训练深度神经网络模型。实验结果表明,针对水稻病害图像,ViT-WGAN-GP模型与GAN、WGAN-GP相比生成图像效果提升显著。使用增强后的水稻病害样本集训练VGG16、ResNet34和GoogLeNet模型,水稻病害识别平均准确率分别达到94.3%,96.2%,97.5%,分别提升了9.7%,2.8%,4.8%。由此可见,该ViT-WGAN-GP模型能生成较为真实的水稻病害图像,且能在小样本集下,较大幅度提高深度神经网络模型的识别准确率。展开更多
With the rapid development of the industrial Internet,the network security environment has become increasingly complex and variable.Intrusion detection,a core technology for ensuring the security of industrial control...With the rapid development of the industrial Internet,the network security environment has become increasingly complex and variable.Intrusion detection,a core technology for ensuring the security of industrial control systems,faces the challenge of unbalanced data samples,particularly the low detection rates for minority class attack samples.Therefore,this paper proposes a data enhancement method for intrusion detection in the industrial Internet based on a Self-Attention Wasserstein Generative Adversarial Network(SA-WGAN)to address the low detection rates of minority class attack samples in unbalanced intrusion detection scenarios.The proposed method integrates a selfattention mechanism with a Wasserstein Generative Adversarial Network(WGAN).The self-attention mechanism automatically learns important features from the input data and assigns different weights to emphasize the key features related to intrusion behaviors,providing strong guidance for subsequent data generation.The WGAN generates new data samples through adversarial training to expand the original dataset.In the SA-WGAN framework,the WGAN directs the data generation process based on the key features extracted by the self-attention mechanism,ensuring that the generated samples exhibit both diversity and similarity to real data.Experimental results demonstrate that the SA-WGAN-based data enhancement method significantly improves detection performance for attack samples from minority classes,addresses issues of insufficient data and category imbalance,and enhances the generalization ability and overall performance of the intrusion detection model.展开更多
文摘针对于当前人脸遮挡修复方法中出现修复图像信息不完整、纹理模糊、产生伪影、细节欠佳以及模型训练不稳定等问题,提出一种基于CSWin-Transformer和WGAN的人脸遮挡修复方法。该方法以Encoder-Decoder结构作为生成器,在生成器中引入CSWin-Transformer Block来精细识别和处理被遮挡的面部区域,以提高处理的针对性和效率,解码器通过跳跃连接与编码器多尺度特征融合,更好学习图像的细节特征,优化最终效果。在判别器中引入Wasserstein距离,来提高模型训练稳定性以及生成图像的真实性,同时在判别器中引入CSWinSelf-Attention,增强判别器对图像全局结构和细节信息的理解。实验结果显示,文章方法在所使用的CelebA的数据集上有良好的修复效果,在峰值信噪比(PSNR)和结构相似性指数(SSIM)指标上与目前一些图像修复方法相比表现更优。 In view of the problems of incomplete repair image information, blurred texture, artifacts, poor details and unstable model training, a face occlusion repair method based on CSWin-Transformer and WGAN is proposed. This method takes Encoder-Decoder, structure as the generator, and introduces CSWin-Transformer Block in the generator to finely identify and process the occluded face areas, so as to improve the pertinacity and efficiency of processing. The decoder integrates with the encoder multi-scale features through jump connection to better learn the detailed features of the image and optimize the final effect. The Wasserstein distance is introduced into the discriminator to improve the stability of the model training and the authenticity of the generated image. Meanwhile, CSWin Self-Attention is introduced in the discriminator to enhance the understanding of the global structure and details of the image. The experimental results show that the method has good repair effect on the data set of CelebA used, and better than some current image repair methods in peak signal-to-noise ratio (PSNR) and structural similarity index (SSIM) indicators.
基金supported by the National Natural Science Foundation of China(62473341)Key Technologies R&D Program of Henan Province(242102211071,252102211086,252102210166).
文摘With the rapid development of the industrial Internet,the network security environment has become increasingly complex and variable.Intrusion detection,a core technology for ensuring the security of industrial control systems,faces the challenge of unbalanced data samples,particularly the low detection rates for minority class attack samples.Therefore,this paper proposes a data enhancement method for intrusion detection in the industrial Internet based on a Self-Attention Wasserstein Generative Adversarial Network(SA-WGAN)to address the low detection rates of minority class attack samples in unbalanced intrusion detection scenarios.The proposed method integrates a selfattention mechanism with a Wasserstein Generative Adversarial Network(WGAN).The self-attention mechanism automatically learns important features from the input data and assigns different weights to emphasize the key features related to intrusion behaviors,providing strong guidance for subsequent data generation.The WGAN generates new data samples through adversarial training to expand the original dataset.In the SA-WGAN framework,the WGAN directs the data generation process based on the key features extracted by the self-attention mechanism,ensuring that the generated samples exhibit both diversity and similarity to real data.Experimental results demonstrate that the SA-WGAN-based data enhancement method significantly improves detection performance for attack samples from minority classes,addresses issues of insufficient data and category imbalance,and enhances the generalization ability and overall performance of the intrusion detection model.
