A Network Garment Style Design System (NGSDS) is proposed to enable the remote style structure drawing design of garment. After the development of the system structure based on network that consists of client end and ...A Network Garment Style Design System (NGSDS) is proposed to enable the remote style structure drawing design of garment. After the development of the system structure based on network that consists of client end and server end at two remote places, a multi-layer part database based on Oracle platform is presented to store information of different parts of garment style. With the acquirement of remote design data at server end using Http technology, the style design is ultimately implemented at the client end using Auto-connecting algorithms. One empirical example is given to show the implementation of the NGSDS.展开更多
A popular TV series in 2012,"Legend of Zhen Huan"reflecting the Harem struggle of Qing Dynasty was broadcasted by several TV stations time after time,obtaining high audience rating.With the improvement of th...A popular TV series in 2012,"Legend of Zhen Huan"reflecting the Harem struggle of Qing Dynasty was broadcasted by several TV stations time after time,obtaining high audience rating.With the improvement of the popularity,the"Zhen Huan’s Mania"phenomenon occurred in many aspects.This thesis will describe the structure of Zhen Huan style,and offer a linguistic account of their popularity in the cyberspace.展开更多
Computer-aided diagnostic systems can assist doctors in diagnosing and treating DR cases more effectively, thereby improving work efficiency, reducing the burden on doctors during examinations, and alleviating problem...Computer-aided diagnostic systems can assist doctors in diagnosing and treating DR cases more effectively, thereby improving work efficiency, reducing the burden on doctors during examinations, and alleviating problems related to uneven distribution of medical resources and shortage of doctors. In this article, we propose a classification method for diabetic retinopathy based on a bilinear multi-attention network. This method uses two backbone networks to extract features, and cross-shares the features using two attention modules to further deepen feature extraction. The non-local attention module is added to address the limitations of traditional convolutional neural networks in capturing global information. By paying attention to highly correlated pathological areas globally, performance improvement can be achieved. We achieved an accuracy of 91.7% on the Messidor dataset.展开更多
生成对抗网络常常被用于图像着色、语义合成、风格迁移等图像转换任务,但现阶段图像生成模型的训练往往依赖于大量配对的数据集,且只能实现两个图像域之间的转换。针对以上问题,提出了一种基于生成对抗网络的时尚内容和风格迁移模型(con...生成对抗网络常常被用于图像着色、语义合成、风格迁移等图像转换任务,但现阶段图像生成模型的训练往往依赖于大量配对的数据集,且只能实现两个图像域之间的转换。针对以上问题,提出了一种基于生成对抗网络的时尚内容和风格迁移模型(content and style transfer based on generative adversarial network,CS-GAN)。该模型利用对比学习框架最大化时尚单品与生成图像之间的互信息,可保证在时尚单品结构不变的前提下实现内容迁移;通过层一致性动态卷积方法,针对不同风格图像自适应地学习风格特征,实现时尚单品任意风格迁移,对输入的时尚单品进行内容特征(如颜色、纹理)和风格特征(如莫奈风、立体派)的融合,实现多个图像域的转换。在公开的时尚数据集上进行对比实验和结果分析,该方法与其他主流方法相比,在图像合成质量、Inception score和FID距离评价指标上均有所提升。展开更多
文摘A Network Garment Style Design System (NGSDS) is proposed to enable the remote style structure drawing design of garment. After the development of the system structure based on network that consists of client end and server end at two remote places, a multi-layer part database based on Oracle platform is presented to store information of different parts of garment style. With the acquirement of remote design data at server end using Http technology, the style design is ultimately implemented at the client end using Auto-connecting algorithms. One empirical example is given to show the implementation of the NGSDS.
文摘A popular TV series in 2012,"Legend of Zhen Huan"reflecting the Harem struggle of Qing Dynasty was broadcasted by several TV stations time after time,obtaining high audience rating.With the improvement of the popularity,the"Zhen Huan’s Mania"phenomenon occurred in many aspects.This thesis will describe the structure of Zhen Huan style,and offer a linguistic account of their popularity in the cyberspace.
文摘Computer-aided diagnostic systems can assist doctors in diagnosing and treating DR cases more effectively, thereby improving work efficiency, reducing the burden on doctors during examinations, and alleviating problems related to uneven distribution of medical resources and shortage of doctors. In this article, we propose a classification method for diabetic retinopathy based on a bilinear multi-attention network. This method uses two backbone networks to extract features, and cross-shares the features using two attention modules to further deepen feature extraction. The non-local attention module is added to address the limitations of traditional convolutional neural networks in capturing global information. By paying attention to highly correlated pathological areas globally, performance improvement can be achieved. We achieved an accuracy of 91.7% on the Messidor dataset.
文摘生成对抗网络常常被用于图像着色、语义合成、风格迁移等图像转换任务,但现阶段图像生成模型的训练往往依赖于大量配对的数据集,且只能实现两个图像域之间的转换。针对以上问题,提出了一种基于生成对抗网络的时尚内容和风格迁移模型(content and style transfer based on generative adversarial network,CS-GAN)。该模型利用对比学习框架最大化时尚单品与生成图像之间的互信息,可保证在时尚单品结构不变的前提下实现内容迁移;通过层一致性动态卷积方法,针对不同风格图像自适应地学习风格特征,实现时尚单品任意风格迁移,对输入的时尚单品进行内容特征(如颜色、纹理)和风格特征(如莫奈风、立体派)的融合,实现多个图像域的转换。在公开的时尚数据集上进行对比实验和结果分析,该方法与其他主流方法相比,在图像合成质量、Inception score和FID距离评价指标上均有所提升。