非径向管板焊接接头的超声检测

对AP1000压水堆核电站安全壳贯穿件的非径向管板焊接接头结构特点进行了介绍,并分析其超声检测工艺的重点和难点。通过建立统一的缺陷定位坐标系并结合被检构件的几何特点进行综合分析,推导得出了h值(缺陷回波最高点距套管外壁的距离)的计算方法。为超声检测对这类焊接接头缺陷的准确识别和定位提供了行之有效的方法,为解决这类焊接接头的超声检测技术受限区提供了行之有效的措施。     

Crosstalk between MicroRNA and Oxidative Stress in Primary Open-Angle Glaucoma

Reactive oxygen species (ROS) plays a key role in the pathogenesis of primary open-angle glaucoma (POAG), a chronic neurodegenerative disease that damages the trabecular meshwork (TM) cells, inducing apoptosis of the retinal ganglion cells (RGC), deteriorating the optic nerve head, and leading to blindness. Aqueous humor (AH) outflow resistance and intraocular pressure (IOP) elevation contribute to disease progression. Nevertheless, despite the existence of pharmacological and surgical treatments, there is room for the development of additional treatment approaches. The following review is aimed at investigating the role of different microRNAs (miRNAs) in the expression of genes and proteins involved in the regulation of inflammatory and degenerative processes, focusing on the delicate balance of synthesis and deposition of extracellular matrix (ECM) regulated by chronic oxidative stress in POAG related tissues. The neutralizing activity of a couple of miRNAs was described, suggesting effective downregulation of pro-inflammatory and pro-fibrotic signaling pathways, including nuclear factor kappa-light-chain-enhancer of activated B cells (NF-kB), transforming growth factor-beta 2 (TGF-β2), Wnt/β-Catenin, and PI3K/AKT. In addition, with regards to the elevated IOP in many POAG patients due to increased outflow resistance, Collagen type I degradation was stimulated by some miRNAs and prevented ECM deposition in TM cells. Mitochondrial dysfunction as a consequence of oxidative stress was suppressed following exposure to different miRNAs. In contrast, increased oxidative damage by inhibiting the mTOR signaling pathway was described as part of the action of selected miRNAs. Summarizing, specific miRNAs may be promising therapeutic targets for lowering or preventing oxidative stress injury in POAG patients.     

堆本体抗震试验动力相似关系推导与验证

为确保堆本体抗震试验中流体对流效应、脉冲效应和堆本体结构响应的准确性，需保证重力、流体与固体惯性力、结构弹性力和结构应变的相似性。本文从固体结构的振动方程、不可压牛顿流体的动力学方程、流固交界面的边界条件和环形柱体域内液体线性晃动的动力学公式出发，基于控制方程的量纲分析法，推导了考虑液体晃动效应的堆本体地震响应动力相似关系。基于上述相似关系建立了堆容器堆内构件和堆容器内自由液面流体域的缩尺模型，通过有限体积法分析堆容器堆内构件原型和缩尺模型中液体的晃动固有频率、晃动波高、压力以及液体晃动对堆容器支承裙的倾覆力矩。结果表明本文动力相似关系具有合理性和准确性，可用于堆本体缩尺模型的抗震试验研究。     

TAS-116, a Well-Tolerated Hsp90 Inhibitor,Prevents the Activation of the NLRP3 Inflammasome in Human Retinal Pigment Epithelial Cells

Chronic inflammation has been associated with several chronic diseases, such as age-related macular degeneration (AMD). The NLRP3 inflammasome is a central proinflammatory signaling complex that triggers caspase-1 activation leading to the maturation of IL-1β. We have previously shown that the inhibition of the chaperone protein, Hsp90, prevents NLRP3 activation in human retinal pigment epithelial (RPE) cells; these are cells which play a central role in the pathogenesis of AMD. In that study, we used a well-known Hsp90 inhibitor geldanamycin, but it cannot be used as a therapy due to its adverse effects, including ocular toxicity. Here, we have tested the effects of a novel Hsp90 inhibitor, TAS-116, on NLRP3 activation using geldanamycin as a reference compound. Using our existing protocol, inflammasome activation was induced in IL-1α-primed ARPE-19 cells with the proteasome and autophagy inhibitors MG-132 and bafilomycin A1, respectively. Intracellular caspase-1 activity was determined using a commercial caspase-1 activity kit and the FLICA assay. The levels of IL-1β were measured from cell culture medium samples by ELISA. Cell viability was monitored by the 3-(4,5-dimethylthiazol-2-yl)-2,5-diphenyltetrazolium bromide (MTT) test and lactate dehydrogenase (LDH) measurements. Our findings show that TAS-116 could prevent the activation of caspase-1, subsequently reducing the release of mature IL-1β. TAS-116 has a better in vitro therapeutic index than geldanamycin. In summary, TAS-116 appears to be a well-tolerated Hsp90 inhibitor, with the capability to prevent the activation of the NLRP3 inflammasome in human RPE cells.     

结合注意力的3D卷积网络脑胶质瘤分割算法

为了提升脑胶质瘤分割精度，提出一种结合注意力机制的3D卷积神经网络算法。输入3个不同尺度的图像块，经过9个卷积层和1个分类层后得到3个不同的分类结果，将分类结果与注意力学习到的权重相乘并逐体素相加得到输出。此外该算法采用了一种混合Dice损失函数与Focal损失函数的超参数损失函数。实验表明，该算法的Dice系数在整体区域、核心区域以及增强区域分别达到了95.31%、80.12%、82.25%。与已有的一种脑胶质瘤分割算法deepmedic相比，整体区域、核心区域以及增强区域的Dice系数分别提升了3%、2%、6%。在脑胶质瘤分割方面，具有重要的临床意义。     

基于卷积神经网络的单目深度估计

针对利用深度学习方法对街道图像进行深度估计，提出采用语义分割的方法解决深度图出现边界模糊等问题；估计深度通过左右视角图生成视差图进行无监督的训练。在网络模型中添加语义分割层，采取多个空洞卷积并行的结构增加感受野，同时减少了图像下采样的次数，降低了由于下采样带来的信息损失，使得的结果更加准确。这也是在深度估计中首次与空洞卷积相结合增加准确率。通过对KITTI街道数据集进行训练，与现有结果相比，除了增加检测准确性，降低错误率之外，使得效果图中的物体更加清晰，并且在效果图中还保留了一些原模型中被忽视掉的细节信息，将原始图像更加完整的表现出来。     

基于机器视觉的运煤车车号识别技术研究

在铁路运煤装车过程中为了快速、准确地识别车号,提出一种基于机器视觉的运煤车车号识别技术。将连通区域提取与投影分割法结合,实现车号的粗定位、细分割,并对图像中的断裂字符进行二次分割,构建了基于BP神经网络的分类模型进行车号识别,提升了煤炭装车的效率和精度。     

R2D2: A scalable deep learning toolkit for medical imaging segmentation

Deep learning has gained a significant popularity in recent years thanks to its tremendous success across a wide range of relevant fields of applications, including medical image analysis domain in particular. Although convolutional neural networks (CNNs) based medical applications have been providing powerful solutions and revolutionizing medicine, efficiently training of CNNs models is a tedious and challenging task. It is a computationally intensive process taking long time and rare system resources, which represents a significant hindrance to scientific research progress. In order to address this challenge, we propose in this article, R2D2, a scalable intuitive deep learning toolkit for medical imaging semantic segmentation. To the best of our knowledge, the present work is the first that aims to tackle this issue by offering a novel distributed versions of two well-known and widely used CNN segmentation architectures [ie, fully convolutional network (FCN) and U-Net]. We introduce the design and the core building blocks of R2D2. We further present and analyze its experimental evaluation results on two different concrete medical imaging segmentation use cases. R2D2 achieves up to 17.5× and 10.4× speedup than single-node based training of U-Net and FCN, respectively, with a negligible, though still unexpected segmentation accuracy loss. R2D2 offers not only an empirical evidence and investigates in-depth the latest published works but also it facilitates and significantly reduces the effort required by researchers to quickly prototype and easily discover cutting-edge CNN configurations and architectures.     

结合深度学习与支持向量机的金属零件识别

目的 在视觉引导的工业机器人自动拾取研究中，关键技术难点之一是机器人抓取目标区域的识别问题。特别是金属零件，其表面的反光、随意摆放时相互遮挡等非结构化因素都给抓取区域的识别带来巨大的挑战。因此，本文提出一种结合深度学习和支持向量机的抓取区域识别方法。方法 分别提取抓取区域的方向梯度直方图（HOG）和局部二进制模式（LBP）特征，利用主成分分析法（PCA）对融合后的特征进行降维，以此来训练支持向量机（SVM）分类器。通过训练Mask R-CNN（regions with convolutional neural network）神经网络完成抓取区域的初步分割。然后利用SVM对Mask R-CNN识别的抓取区域进行二次分类，完成对干扰区域的剔除。最后计算掩码完成实例分割，以此达到对抓取区域的精确识别。结果 对于随机摆放的铜质金属零件，本文算法与单一的Mask R-CNN及多特征融合的SVM算法就识别准确率、错检率、漏检率3个指标进行了比较，结果表明本文算法在识别准确率上较Mask R-CNN和SVM算法分别提高了7%和25%，同时有效降低了错检率与漏检率。结论 本文算法结合了Mask R-CNN与SVM两种方法，对于反光和遮挡情况具有一定的鲁棒性，同时有效地提升了目标识别的准确率。     

智能跟踪算法在采砂监管中的应用研究

受利益的驱动,长江流域非法采砂行为屡禁不止,给采砂有效监管造成很大的困扰。采砂监管水平的提高需要高效准确的跟踪识别技术作支撑。提出了一种基于相关滤波器的可变跟踪框智能算法。该算法可以准确快速地跟踪江河湖面上的非法采砂船,计算量少,鲁棒性高,对硬件的要求并不严苛,可以自由应用于从低性能到高性能的各种设备。基于该算法开发了非法采砂船追踪器,再配合百度深度学习平台Easy DL后,可构成一套采砂船监控系统,能够实现对长江水域非法采砂船24 h的有效监控,从而大大提高监控非法采砂行为的效率。