2015年6月9日,2015 e Net16论坛在北京798艺术区尤伦斯当代艺术中心举行。来自母婴、生鲜电商、在线教育、数字营销、云计算等互联网热门创新领域的十五家成长型企业代表、以及部分业内专家围绕＂电商如何联接家庭厨房＂、＂寻找家庭服...2015年6月9日,2015 e Net16论坛在北京798艺术区尤伦斯当代艺术中心举行。来自母婴、生鲜电商、在线教育、数字营销、云计算等互联网热门创新领域的十五家成长型企业代表、以及部分业内专家围绕＂电商如何联接家庭厨房＂、＂寻找家庭服务产业的风口＂、＂从蛮荒到文明,在线教育的品牌成长之路＂三轮主题对话交流,就2015上半年互联网行业的整体发展和显著特点进行了深入的探讨和总结,阐述了各自企业的价值主张。展开更多
Background Bone morphogenetic protein （BMP）-2, alkaline phosphatase （ALP）, and collagen type I are known to play a critical role in the process of bone remodeling. However, the relationship between mechanical stra...Background Bone morphogenetic protein （BMP）-2, alkaline phosphatase （ALP）, and collagen type I are known to play a critical role in the process of bone remodeling. However, the relationship between mechanical strain and the expression of BMP-2, ALP, and COL-I in osteoblasts was still unknown. The purpose of this study was to investigate the effects of different magnitudes of mechanical strain on osteoblast morphology and on the expression of BMP-2, ALP, and COL-I. Methods Osteoblast-like cells were flexed at four deformation rates （0, 6%, 12%, and 18% elongation）. The expression of BMP-2 mRNA, ALP, and COL-I in osteoblast-like cells were determined by real-time quantitative reverse transcription polymerase chain reaction, respectively. The results were subjected to analysis of variance （ANOVA） using SPSS 13.0 statistical software. Results The cells changed to fusiform and grew in the direction of the applied strain after the mechanical strain was loaded. Expression level of the BMP-2, ALP, and COL-I increased magnitude-dependently with mechanical loading in the experimental groups, and the 12% elongation group had the highest expression （P 〈0.05）. Conclusion Mechanical strain can induce morphological change and a magnitude-dependent increase in the expression of BMP-2, ALP, and COL-I mRNA in osteoblast-like cells, which might influence bone remodeling in orthodontic treatment.展开更多
针对现有集装箱编号识别算法结构复杂以及难以应用在低成本设备上的问题,文中提出一种基于深度学习的轻量型算法YOLOv4-GSE(You Only Look Once version 4-Ghost-SPPFA-Effective),利用Ghost-Enet网络替代YOLOv4的主干特征提取网络,使用...针对现有集装箱编号识别算法结构复杂以及难以应用在低成本设备上的问题,文中提出一种基于深度学习的轻量型算法YOLOv4-GSE(You Only Look Once version 4-Ghost-SPPFA-Effective),利用Ghost-Enet网络替代YOLOv4的主干特征提取网络,使用Ghost卷积替换所有3×3卷积方式来削减模型的参数量。在加强特征提取部分,提出一种改进后的SPPFA(Spatial Pyramid Pooling with Feature Aggregation)模块来解决由于连续最大池化操作造成的信息丢失问题。添加CBAM(Convolutional Block Attention Module)注意力机制模块对不同通道和空间进行权重分析,增强模型的特征提取能力。相较于YOLOv4,所提算法在集装箱数据集上的mAP(mean Average Precision)值提升了1.02%,参数量减少了91.95%,FLOPs(Floating-point Operations Per Second)减少了94.62%。展开更多
文摘2015年6月9日,2015 e Net16论坛在北京798艺术区尤伦斯当代艺术中心举行。来自母婴、生鲜电商、在线教育、数字营销、云计算等互联网热门创新领域的十五家成长型企业代表、以及部分业内专家围绕＂电商如何联接家庭厨房＂、＂寻找家庭服务产业的风口＂、＂从蛮荒到文明,在线教育的品牌成长之路＂三轮主题对话交流,就2015上半年互联网行业的整体发展和显著特点进行了深入的探讨和总结,阐述了各自企业的价值主张。
文摘Background Bone morphogenetic protein （BMP）-2, alkaline phosphatase （ALP）, and collagen type I are known to play a critical role in the process of bone remodeling. However, the relationship between mechanical strain and the expression of BMP-2, ALP, and COL-I in osteoblasts was still unknown. The purpose of this study was to investigate the effects of different magnitudes of mechanical strain on osteoblast morphology and on the expression of BMP-2, ALP, and COL-I. Methods Osteoblast-like cells were flexed at four deformation rates （0, 6%, 12%, and 18% elongation）. The expression of BMP-2 mRNA, ALP, and COL-I in osteoblast-like cells were determined by real-time quantitative reverse transcription polymerase chain reaction, respectively. The results were subjected to analysis of variance （ANOVA） using SPSS 13.0 statistical software. Results The cells changed to fusiform and grew in the direction of the applied strain after the mechanical strain was loaded. Expression level of the BMP-2, ALP, and COL-I increased magnitude-dependently with mechanical loading in the experimental groups, and the 12% elongation group had the highest expression （P 〈0.05）. Conclusion Mechanical strain can induce morphological change and a magnitude-dependent increase in the expression of BMP-2, ALP, and COL-I mRNA in osteoblast-like cells, which might influence bone remodeling in orthodontic treatment.
文摘针对现有集装箱编号识别算法结构复杂以及难以应用在低成本设备上的问题,文中提出一种基于深度学习的轻量型算法YOLOv4-GSE(You Only Look Once version 4-Ghost-SPPFA-Effective),利用Ghost-Enet网络替代YOLOv4的主干特征提取网络,使用Ghost卷积替换所有3×3卷积方式来削减模型的参数量。在加强特征提取部分,提出一种改进后的SPPFA(Spatial Pyramid Pooling with Feature Aggregation)模块来解决由于连续最大池化操作造成的信息丢失问题。添加CBAM(Convolutional Block Attention Module)注意力机制模块对不同通道和空间进行权重分析,增强模型的特征提取能力。相较于YOLOv4,所提算法在集装箱数据集上的mAP(mean Average Precision)值提升了1.02%,参数量减少了91.95%,FLOPs(Floating-point Operations Per Second)减少了94.62%。
