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.     

融合语义标签和噪声先验的图像生成

针对现有生成模型难以直接从复杂语义标签生成高分辨率图像的问题,提出了融合语义标签和噪声先验的生成对抗网络(SLNP-GAN)。首先,直接输入语义标签(包含形状、位置和类别等信息),使用全局生成器对其进行编码,并结合噪声先验来学习粗粒度的全局属性,初步合成低分辨率图像;然后,基于注意力机制,使用局部细化生成器来查询低分辨率图像子区域对应的高分辨率子标签,获取细粒度信息,从而生成纹理清晰的复杂图像;最后,采用改进的引入动量的Adam算法(AMM)算法来优化对抗训练。实验结果表明,与现有方法text2img相比,所提方法的像素精确度(PA)在COCO_Stuff和ADE20K数据集上分别提高了23.73%和11.09%;相较于Adam算法,AMM算法收敛速度提升了约一倍,且损失值波幅较小。可见,SLNP-GAN能高效地获取全局特征和局部纹理,生成细粒度、高质量的图像。     

Telesurgery QoS improvement over SDN based on a Type‐2 fuzzy system and enhanced cuckoo optimization algorithm

Telemedicine is a new area based on the information and communication technology for collecting, storing, organizing, retrieving and exchanging medical information. One of the most important applications of telemedicine is indeed telesurgery in which an efficient telecommunication infrastructure between the surgery room and remote surgeons need to be established. One of the most important issues to be tackled in telesurgery is to find favorable links for routing as well as providing high Quality of Service (QoS). In this paper, an efficient model based on the hybridization of Type‐2 Fuzzy System (T2FS) and Cuckoo Optimization Algorithm (COA) over the Software Defined Networks (SDN) is proposed in order to achieve optimal and reliable routes for telesurgery application. Using T2FS, the fitness of the links is determined; then, a COA is conducted over the Constraint Shortest Path (CSP) problem to find the best routes. Delay is considered as a CSP problem which is satisfied by trying to find the paths with minimum cost. Due to the NP‐completeness of the CSP problem, an Enhanced COA (so‐called E‐COA) is proposed and utilized as a metaheuristic solver. To the best of our knowledge, this paper is the first SDN‐based communication model that applies both T2FS and E‐COA for assigning proper costs to the network's links, and solves the consequence CSP problem according to the QoS requirement for telesurgery. The model also recognizes and preserves the second‐best routes in order to keep the reliability for such a critical application. In addition to the simulations, the performance evaluation is also conducted on a real experimental scenario. Many comparisons are carried out between the proposed model and other conventional methods, and the evaluation study shows the superiority of the proposed model on all the three QoS‐related metrics, i.e. average end‐to‐end delay, packet loss ratio and PSNR.     

云计算在智慧医院信息服务中的应用

随着医疗服务体系的发展和进步,越来越多的医院开始将云计算技术引入信息服务系统,使医院信息化建设得到了迅速发展。文章分析了云计算技术的特点及其在智能医院信息服务中的应用和影响,以期促进该技术在未来的发展和推广。     

Design of frameworks for self-adaptive service-oriented applications: A systematic analysis

Self-adaptive service-oriented Applications (Self-Apps) must be able to understand themselves or the environment in which they are executed, and propose solutions to meet changing conditions. The development of these applications is not a trivial task, since it encompasses issues from different research areas. Despite the importance of frameworks for Self-Apps, there is a lack of comprehensive analysis of how the design of such applications is performed, and regarding the standardization of concepts and coverage of minimum requirements for Self-Apps. The main contribution of this article is to present this comprehensive analysis, providing the state of the art for this subject. This analysis was built through a Systematic Mapping Study, based on a total of 65 studies, from which we identify the main attributes for Quality of Service (QoS), search strategies, and service management strategies employed in the design of frameworks for Self-Apps. The main aspects of requirements involved in the design of Self-Apps were pointed out to stakeholders. For example, these applications must implement a method for evaluation of QoS based on metrics. We also put forward the S-Frame, a modular solution that brings together the main features for the design of Self-Apps, and describe the main challenges concerning these applications.     

基于NB分类器重访概率预测的Web缓存替换策略

Web缓存是用来解决网络访问延迟和网络拥塞问题，缓存替换策略直接影响缓存的命中率。为此，提出一种朴素贝叶斯（NB）分类器重访概率预测的Web缓存替换策略；根据用户之前访问日志，通过分区操作提取多项特征来表示每次访问的对象，并构建特征数据集；训练NB分类器，用来确定缓存中对象被再次访问的概率，为对象分配权重；结合LRU策略来合理删除一些对象。仿真结果表明，提出的策略在保证较高命中率的同时有效降低了执行时间。     

大坝服役非概率可靠性分析方法

针对大坝服役可靠性影响因子具有非概率不确定特征的情况,建立了定量刻画大坝服役可靠性非概率影响因子的Info-gap模型,运用Info-gap决策理论和体积比非概率可靠性度量方法,提出了大坝服役非概率可靠性分析方法,拟定了大坝服役非概率目标可靠度。应用提出的非概率方法分析了某大坝结构服役可靠性,各功能模式可靠性分析结果说明了该方法的工程适用性,同时也验证了所拟定大坝服役非概率目标可靠度的合理性。     

Economic Evaluation of Power System Flexibility by Commercial Air Conditioner Control with Large Penetration of Photovoltaic Generation

High penetration of intermittent renewable energy such as photovoltaic (PV) and wind power could cause shortage of power system flexibility. Demand response is expected to help supply ancillary service instead of the conventional power plant. Commercial air conditioners are a promising responsive load for demand response because they account for a large proportion of power consumption in the power system. We calculate a system operation cost and hourly operation pattern of each power plant by using the optimal power generation model considering flexibility supply from controlling commercial air conditioner. We obtained the following results as an effect of commercial air conditioner control. (1) The power generation of oil fired power plants decreases at peak time and annual fuel cost of oil fired plant is reduced by approximately 30% at most in Kanto area. (2) The percentage of rated operation mode of LNG combined cycle plants increases. (3) Curtailed energy rate of PV decreases because a power storage amount by pumped hydropower generation increases. (4) Required battery capacity to reduce PV curtailed amount decreased by combining battery energy storage system in case of high penetration of PV.     

基于符号模型检测的Web服务组合形式化验证

随着经济的发展和市场竞争的加剧,企业必须能够快速且准确地满足市场和用户的各种需求。Web服务组合正是由于单个Web服务不能满足企业及用户的需求而产生的一种技术,而如何确保组合的正确性来实现服务增值是一个尚未完全解决的问题。针对此问题,提出了一种基于符号模型检测器NuSMV对Web服务组合进行验证的方法,并提出了基于消息会话的Web服务有限状态自动机的形式化定义。最后实例验证了Web服务组合交互的正确性和有无死锁状态现象,进一步证明了方法的可行性。