波动方程all-at-once系统的快速α循环绝对值预处理

为了加快预处理MINRES方法求解波动方程all-at-once系统的收敛速度,基于绝对值预处理子和块状三对角Toeplitz预处理子,提出一种新的α循环绝对值预处理子。理论上证明了预处理矩阵可近似分裂成正交矩阵与低秩矩阵的和,且其特征值聚集在±1附近,保证了预处理MINRES方法的快速收敛性质。数值实验结果进一步表明了新预处理子的有效性。

基于MMC2107 OnCE的Flash编程技术

阐述了如何通过Motorola32位MCUMMC2107的OnCE接口,实现对MMC2107的片外Flash存储器的编程。详细介绍了OnCE模块的实现原理、内部工作机制、硬件系统设计、软件编程方法及程序接口。并且提出了OnCE简易硬件接口的原理设计、编程模型、软件接口划分、操作流程,以及提高Flash编程速度的改进方案,探讨其中一些关键技术和实现细节。

Study on the Optimum Moisture Amount Added to Brown Rice once during the Wet Conditioning

In the experiment, the brown rice whose moisture content was 12.5% was used as raw material. The brown rice was grouped, then moisturized differently and milled. While milling, the energy consumption, the rate of broken rice and the crack rate were tested. It is confirmed that the stress crack owing to the moisture added to the brown rice can be avoided when the moisture amount added once is limited to no more than 1.5%. It is also proved that the energy consumption can be reduced, the yielding rate of rice can be increased and that the quality of rice can be improved.

基于You Only Look Once v2优化算法的车辆实时检测

针对基于You Only Look Once v2算法的目标检测存在精度低及稳健性差的问题,提出一种车辆目标实时检测的You Only Look Once v2优化算法;该算法以You Only Look Once v2算法为基础,通过增加网络深度,增强特征提取能力,同时,通过添加残差模块,解决网络深度增加带来的梯度消失或弥散问题;该方法将网络结构中低层特征与高层特征进行融合,提升对小目标车辆的检测精度。结果表明,通过在KITTI数据集上进行测试,优化后的算法在检测速度不变的情况下,提高了车辆目标检测精度,平均精度达到0.94,同时提升了小目标检测的准确性。

A dynamic model of once-daily 5-aminosalicylic acid predicts clinical efficacy

New once daily mesalamine formulations may improve adherence to medication usage.Response to Asacol and other forms of 5-aminosalicyclic acid(5-ASA)is better correlated with tissue concentrations and best predicted by concentrations of the drug within the lumen of the colon.Our group used computer simulation to predict colonic 5-ASA levels after Asacol administration.In our study,the model simulated Asacol distribution in the healthy colon,and during quiescent and active ulcerative colitis.An Asacol dosage of 800 mg,threetimes a day,was compared to 2400 mg given once a day.Under ideal conditions,the predicted maximum drug in the total colon and individual colonic segments over 100 d differed by less than 3%between single and multiple doses.Despite changes in motility and defection rates,the predicted maximum and average 5-ASA concentrations in the total colon and individual colonic segments differed by less than 10%between dosing regimens.Asymmetric distribution of 5-ASA in the colon was influenced by frequency of bowel movements and colonic transit rate.In active colitis,sigmoid 5-ASA concentration becomes negligible.Our model supports once daily administration of Asacol as standard treatment for ulcerative colitis.

The Antifreeze Critical Strength of Low-temperature Concrete Effected by Index

The antifreeze critical strength and the pre-curing time of low-temperature concrete were studied by means of guaranteed rate of compressive strength and antifreeze performance for the structural safety requirement of concrete engineering,suffering once freeze damage under air environment.It is shown that the antifreeze critical strength is 3.7-4.4MPa,pre-curing time is 18-32 h by guaranteed rate of compressive strength,and the antifreeze critical strength is 3.7-4.4MPa,pre-curing time is 18-32 h by guaranteed rate of antifreeze performance.It can be found that the method of guaranteed rate of compressive strength is sensitive to the defect which generated by freeze damage in the concrete interior.The method is fit to evaluate the antifreeze critical strength of low-temperature concrete.

A Three-layered Contrasting Analysis of the Theme in Once more to the lake

Once more to the lake is a representative work of the American Writer, Elwyn Brooks White, and has been acclaimed by people in the world. The paper offers a three-layered contrasting analysis of the essay. By detailed analysis of the contrasts running through the whole essay, the author finds that those contrasts are symbol of White's internal struggle and reflection for life.On the foundation of the three-layered contrasts, the paper presents the theme analysis and clarifies the content of the essay.

An Only-Once-Sorting Algorithm

This paper provides a new sorting algorithm called 'Only-Once-Sorting' algorithm a mathemati cal formula,this algorithm can put elements in the positions they should be stored only once,then compacts them.The algorithm completes sorting a sequence of n elements in a calculation time of O(n ).

Stage Predictions of Landslide and Subsidence from an Once-Through Cycle

In this paper both processes of landslide and subsidence are considered to be limited systems. Each of these systems in nature might be regarded as an organism. Generally their lifespan must develop with common ecological characteristics, including several evolutional stages, such as initiation, growth, maturation, decline and death. Among these stages, maturation is emphasized so as to find the occurring or thriving date of both systems. An once-through cycle of both landslide and subsidence is established and is accurately predicted by a developed, mathematic model of the Poisson cycle. The Weibull distribution is cited for a landslide example. Both fundamentals are discussed. Stage predictions of landslide and subsidence are performed for several examples. Back analysis of landslides that have already happened are studied with the same model. And when compared with results from the biological mathematic model and with practical results, it is found that they correspond. Stage prediction of subsidences is also researched by the principle of the Poisson cycle.

一种基于多尺度的目标检测锚点构造方法

目标检测是计算机视觉领域的研究热点和基础任务,其中基于锚点(Anchor)的目标检测已在众多领域得到广泛应用。当前锚点选取方法主要面临两个问题:基于特定数据集的先验取值尺寸固定、面对不同场景泛化能力弱。计算锚框的无监督K-means算法,受初始值影响较大,对目标尺寸较单一的数据集聚类产生的锚点差异较小,无法充分体现网络多尺度输出的特点。针对上述问题,本文提出一种基于多尺度的目标检测锚点构造方法(multi-scale-anchor,MSA),将聚类产生的锚点根据数据集本身的特性进行尺度的缩放和拉伸,优化的锚点即保留原数据集的特点也体现了模型多尺度的优势。另外,本方法应用在训练的预处理阶段,不增加模型推理时间。最后,选取单阶段主流算法YOLO(You Only Look Once),在多个不同场景的红外或工业场景数据集上进行丰富的实验。结果表明,多尺度锚点优化方法MSA能显著提高小样本场景的检测精度。

Safety and Efficacy of Once-Daily Administration of Mycophenolate Mofetil in Kidney Transplant Patients—A New Treatment Option for Non-Adherence

Non-adherence to immunosuppressive therapy is associated with reduced graft survival. Fifteen long-term follow up kidney transplant patients with stable post-transplant clinical courses were enrolled in the study. All patients were prescribed to tacrolimus prolonged release, mycophenolate mofetil (MMF) and corticosteroid at the time of enrolment. Twice-daily administration of MMF was then converted to a single daily dose. As a result, all immunosuppressive agents were administered simultaneously at one time in the morning. The daily total doses of the three immunosuppressants were identical during the study period. No acute rejection or adverse event was observed during the study period. Blood urea nitrogen and estimated glomerular filtration rate did not change significantly after conversion to once-daily administration. The blood tacrolimus trough level and the Mycophenolic acid (MPA) trough level also did not change significantly after conversion to once-daily administration. Interestingly, the meanmedian MPA trough level remained >3 μg/ml even after conversion to once-daily administration. A sufficient plasma MPA level was maintained after conversion to once-daily administration, and no acute rejection was observed during the study period. To our knowledge, this study is the first to report that the plasma MPA concentration can be maintained after once-daily administration of MMF in long-term kidney transplant patients. Once-daily administration of immunosuppressive agents may improve long-term graft survival because of better treatment adherence due to the reduced dosing frequency. The safety or efficacy of conversion to once-daily administration of MMF should be evaluated in a future randomized controlled large-scale clinical study.

基于改进YOLOv8的嵌入式道路裂缝检测算法

在边缘端设备部署YOLOv8L模型进行道路裂缝检测可以实现较高的精度,但难以保证实时检测。针对此问题,提出一种可部署到边缘计算设备Jetson AGX Xavier上的基于改进YOLOv8模型的目标检测算法。首先,利用部分卷积设计Faster Block结构以替换YOLOv8 C2f模块中的Bottleneck结构,并将改进后的C2f模块记为C2f-Faster;其次,在YOLOv8主干网络中的每个C2f-Faster模块之后接一个SE(Squeeze-and-Excitation)通道注意力层,进一步提高检测的精度。在开源道路损害数据集RDD20(Road Damage Detection 20)上的实验结果表明:所提方法的平均F1得分为0.573,每秒检测帧数(FPS)为47,模型大小为55.5MB,相较于GRDDC2020(GlobalRoadDamageDetection Challenge 2020)的SOTA(State-Of-The-Art)模型,F1得分提高了0.8个百分点,FPS提高了291.7%,模型大小减小了41.8%,实现了在边缘设备上对道路裂缝实时且准确的检测。

基于改进YOLO v5算法的道路小目标检测

为解决交通道路小目标检测难度大、精度低,容易出现错检漏检的问题,提出一种基于YOLO v5(you only look once v5)算法的多尺度特征融合目标检测改进算法。首先,增加小目标检测头用于适应小目标尺寸,缓解漏检情况。然后,引入可变形卷积网络v2(deformable convolutional networks V2,DCN V2)提高模型对运动中小目标的学习能力;同时,增加上下文增强模块,提升对远距离小目标的识别能力。最后,在替换损失函数、提高边界框定位精度的同时,使用空间金字塔池化和上下文空间金字塔卷积分组模块,提高网络的感受野和特征表达能力。实验结果表明,所提算法在KITTI数据集小目标类别上平均识别精度达到了95.2%,相较于原始YOLO v5,算法总体平均识别精度提升了2.7%,对小目标的检测效果更佳,平均识别精度提升了3.1%,证明所提算法在道路小目标检测方面的有效性。

Osteonecrosis of the jaw in a patient taking once-yearly infusion of zoledronic acid for osteopenia

Osteonecrosis of the jaw (ONJ) is an adverse effect of nitrogen-containing bisphosphonates. Advancing age, intravenous administration of zoledronic acid (ZOL), history of dento-alveolar surgery, and concomitant systemic diseases such as diabetes are known as risk factors for developing ONJ. However, despite numerous studies, the exact pathophysiology remains unclear and management strategies are largely anecdotal. Once-yearly intravenously administered 5 mg ZOL was approved by the US Food and Drug Administration in 2007 for the treatment of osteoporosis and its efficacy with 3 year-regimen had been recently proven in preventing new clinical fracture. Although occurrences of ONJ have been reported to be rare with this drug administration, available data is very limited and long-term outcomes are lacking. We present a case of ONJ identified in an osteopenic patient with an intermittent but long standing sore mouth related to exposed mandibular bone. Once-yearly infusion of zoledronic acid used in the treatment of osteopenia may contribute to the spontaneous development of ONJ, especially in those presenting with multiple comorbidity factors. This report suggests the importance of health care professionals keeping abreast of new developments in this area and providing appropriate information to their patients.

基于图像低维特征融合的航拍小目标检测模型

针对无人机航拍图像目标检测中视野变化大、时空信息复杂等问题,文中基于YOLOv5(You Only Look Once Version5)架构,提出基于图像低维特征融合的航拍小目标检测模型.引入CA(Coordinate Attention),改进MobileNetV3的反转残差块,增加图像空间维度信息的同时降低模型参数量.改进YOLOv5特征金字塔网络结构,融合浅层网络中的特征图,增加模型对图像低维有效信息的表达能力,进而提升小目标检测精度.同时为了降低航拍图像中复杂背景带来的干扰,引入无参平均注意力模块,同时关注图像的空间注意力与通道注意力;引入VariFocal Loss,降低负样本在训练过程中的权重占比.在VisDrone数据集上的实验验证文中模型的有效性,该模型在有效提升检测精度的同时明显降低复杂度.

棕榈酸帕利哌酮注射液治疗精神分裂症患者内分泌代谢的影响观察

目的:评估精神分裂症患者接受棕榈酸帕利哌酮注射液1个月剂型(once-monthly paliperidone palmitate,PP1M)治疗后内分泌、代谢情况。方法:对51例口服利培酮治疗有效的精神分裂症患者换用PP1M治疗后,于基线及治疗后1个月、3个月、6个月、12个月监测内分泌、代谢等相关指标,于基线及治疗后12个月对患者进行个体和社会功能量表(personal and social performance scale,PSP)评估。结果:治疗12个月时PSP因子分对社会有益的活动、扰乱及攻击行为两项与基线相比,差异具有统计学意义(t=-2.07、3.71,P均<0.05);重复测量方差分析显示,不同时间点,患者体质量指数、腰围、血糖、血脂等指标差异均无统计学意义(P>0.05);而在血清泌乳素(prolactin,PRL)上,差异具有统计学意义(F=3.021,P<0.05),且各个时间点血清泌乳素浓度均高于基线(P均<0.05)。结论:PP1M对代谢影响较小,且能改善患者社会功能,但会导致血清PRL水平增高,需引起关注。

直流余热锅炉蒸发段长度和位置变动特性的仿真分析

在直流余热锅炉变工况运行过程中蒸发段长度和位置的改变会造成锅炉管壁温度分布、工质出口参数等产生较大的波动,影响锅炉安全运行。以某立式直流余热锅炉为研究对象,建立其动态仿真模型,研究锅炉烟气流量、入口烟温、高压给水流量和给水温度分别发生阶跃扰动后蒸发段长度及相变点位置的变化规律。仿真结果表明:余热锅炉在设计工况下运行,锅炉入口烟温对锅炉蒸发段长度的影响程度最大,入口烟温阶跃升高10℃后,蒸发段迅速缩短,相变点均向锅炉管内工质入口方向移动,蒸发段长度和相变点位置恢复稳定所需时间相对较短;烟气流量阶跃升高5%后,蒸发段长度迅速缩短而后缓慢增长,沸腾点向管内工质入口方向移动距离大于蒸干点,恢复稳定所需时间较长;给水流量阶跃升高5%后,蒸发段长度迅速增长而后缓慢缩短,蒸干点向管内工质出口方向移动距离大于沸腾点,恢复稳定所需时间最长;给水温度阶跃变化20℃对蒸发段长度和相变点位置的影响很小。

琴徽数理的早期应用

汉代时七弦琴已定型,汉魏之际琴面即饰有13徽,间距不一的琴徽暗含立体属性与等分原理。琴徽虽出于汉魏之际,但其一弦上的等分数理渊源甚久,可上溯至商编铙与西周甬钟。“一弦均钟”可能是商周真实存在的、乐师为乐钟作音列设置的准器,也可能不以实物存在,而只是口传心授、隐藏于思维之中的虚拟数理。曾侯乙墓的五弦均钟与两周之际钮钟正鼓音列上的五声,表明两周之际既是五弦均钟应用时间的上限,又是“一弦均钟”应用时间的下限。由于先秦古琴形制尚未统一,钟琴两类乐器在数理应用上存在时间差。从一弦等分节点到商周钟铙四声,从五弦等分节点到东周编钟七声乃至旋宫,钟乐的辉煌始终依赖弦上节点与神瞽们数理认知的支撑。这展现出琴徽使用之前琴徽数理的早期应用和历史贡献。

基于ID-YOLO的数字仪表检测方法

当前针对数字式仪表检测算法在边缘设备具有实时性差、泛化性差的问题,对此提出一种采用ID-YOLO(instrument detection-you only look once)模型的变电站数字仪表检测识别方法。所提算法以YOLOv5模型为基础,首先设计轻量骨干网络(light weight-YOLO, LW-YOLO)提取图像特征,降低网络参数,提高检测实时性;然后设计了一种双级路由注意力模块(bi-level routing attention moudle, BRAM),提高网络对小数点的检测精度以及网络的鲁棒性和泛化性;最后,引入损失函数α-IoU,通过设定不同的可调节参数α数值得到更准确的真实框与预测框的交并比计算,可以提高模型的检测精度。结果表明:相比于其他基于深度学习的数字仪表检测识别方法,所提方法在不同显示方式的数字仪表识别任务上具有更好的准确性和泛化性,而且可以在检测准确率领先的情况下,将模型在边缘设备上的检测速度从6.87帧/s提升至8.77帧/s,其实时性和检测精度均能够满足实际变电站智能数据采集、检测识别的工程需要。

Misp-YOLO:加油站场景目标检测

针对Yolov3-Tiny算法在加油站监控场景检测时由于数据特征提取不充分而导致检测精度低、漏检率高等问题,提出一种基于加油站场景的Misp-YOLO(You Only Look Once)目标检测算法。首先引入Mosaic数据增强算法,使图片包含更多特征信息;其次使用InceptionV2和PSConv(Poly-Scale Convolution)多尺度特征提取方法提升网络多尺度预测能力;最后结合scSE(Concurrent Spatial and Channel ‘Squeeze&Excitation’)注意力机制,重构主干网络输出特征。实验结果证明该算法具有较高检测准确度,并且检测速度满足实际需求。优化后的算法性能得到极大提升,可推广应用于其他目标检测中。