WavNet — Visual saliency detection using Discrete Wavelet Convolutional Neural Network

In the recent advancements in image and video analysis, the detection of salient regions in the image becomes the initial step. This plays a crucial role in deciding the performance of such algorithms. In this work, a Multi-Resolution Feature Extraction (MRFE) technique that makes use of Discrete Wavelet Convolutional Neural Network (DWCNN) for generating features is employed. An Enhanced Feature Extraction (EFE) module extracts additional features from the high level features of the DWCNN, which are used to frame both channel as well as spatial attention models for yielding contextual attention maps. A new hybrid loss function is also proposed, which is a combination of Balanced Cross Entropy (BCE) loss and Edge based Structural Similarity (ESSIM) loss that effectively identifies and segments the salient regions with clear boundaries. The method is tested exhaustively with five different benchmark datasets and is proved superior to the existing state-of-the-art methods with a minimum Mean Absolute error (MAE) of 0.03 and F-measure of 0.956.     

Influence of temperature on acoustic emission source location accuracy in underground structure

An acoustic emission (AE) experiment was carried out to explore the AE location accuracy influenced by temperature. A hollow hemispherical specimen was used to simulate common underground structures. In the process of heating with the flame, the pulse signal of constant frequency was stimulated as an AE source. Then AE signals received by each sensor were collected and used for comparing localization accuracy at different temperatures. Results show that location errors of AE keep the same phenomenon in the early and middle heating stages. In the later stage of heating, location errors of AE increase sharply due to the appearance of cracks. This provides some beneficial suggestions on decreasing location errors of structural cracks caused by temperature and improves the ability of underground structure disaster prevention and control.     

基于测绘地理信息技术的涉地税源风险识别模式应用研究

为解决沧州市涉地税源管理手段单一等问题,文中提出以测绘地理信息技术为支撑,通过对多源数据组织管理,建立风险分析指标,采用多种GIS空间分析方法构建了涉地风险税源识别模式并在沧州进行应用。结果表明:该模式能有效识别多种涉地风险,为涉地税源管理提供了支撑。     

Structural silicone sealants after exposure to laboratory test for durability assessment

During the service life of structural sealant glazing (SSG) facades, the load-bearing capacity of the silicone bonds needs to be guaranteed. Laboratory tests can assess the durability of SSG-systems based on mechanical characteristics of the bond after simultaneous exposure to both climatic and mechanical loads. This article studies how the material characteristics of two common structural sealants are affected by laboratory and field exposure. Dynamic mechanical analysis (DMA) confirms a reduction in the dynamic modulus of exposed silicone samples. Results from thermogravimetric analysis, Fourier-transform infrared spectroscopy, differential scanning calorimetry, and small-angle X-ray scattering/wide-angle X-ray scattering show differences between the two sealants and indicate no/minor changes in the composition and morphology of the laboratory and field exposed sealants. Mechanical characterization methods, such as DMA, and tensile and shear testing of the structural bond, are shown to be sensitive toward the combined climatic and mechanical loadings, and are hence suitable for studying degradation mechanisms of structural sealants.     

地表径流速度对城市内涝影响规律

为分析地表径流速度对城市内涝的影响,采用一维地下排水管网与二维城区地形的动态耦合模型,选取大连市某排水区块作为研究区域,设置4种地表径流速度10种设计降雨场景,模拟分析在不同重现期设计降雨及不同地表径流速度下研究区的内涝积水特性。结果表明:随着地表径流速度降低,管道排水压力变小,管道排水达标率最高可提升48.05%,且降雨重现期越短,地表径流流速对管道排水压力的削减效果越明显;地表径流流速对检查井溢流量影响显著,随着地表径流速度降低,检查井溢流量峰值最高可减小2 750 m~3,峰现时间最长可滞后56 min,同时随着降雨重现期增长,地表径流流速对检查井溢流量的削减效果减弱;研究区低、高风险区淹没面积随地表径流速度降低,最高可分别减小1.64万、8.37万m~2,但中风险区淹没面积变化反复。     

关于COVID-19、膜脂质与公众健康（网络首发、推荐阅读）

冠状病毒具有一层脂质膜。虽然复制需要劫持宿主的RNA工具来合成病毒体蛋白，但必须将其包裹在脂质膜中，促其萌生以扩展感染。最近研究表明，某些必需脂肪酸可以抑制其复制活性。脂质膜通常被认为是水溶物的脂肪屏障，但它对细胞和亚细胞的功能是高度有序和组分特异性的，其对病毒外壳可能也有最佳的特异性。虽然复制中DNA、RNA和蛋白质组成不受饮食影响，但脂质膜受其影响。此外，自1960年代以来，人们就知道男性对这些必需脂肪酸和膜完整性不足的敏感性高于女性。有证据表明，花生四烯酸和二十二碳六烯酸具有抗病毒、免疫、抗炎、控制血压和消退素活性，因此，迫切需要考察它们在Covid-19预防和治疗中的地位，也需重新评估现行的膳食指导。当前，有关脑、神经、血管和免疫系统等富含膜系统对脂质需求还尚未被详细认识。毫无疑问，这些脂质在几百万年来塑造人类基因组方面具有重要意义，因此，如果这些膜脂质失衡将会使人类面临机体紊乱和感染风险，且男性比女性风险更大。     

Data requirements for improving the Quantitative Risk Assessment of liquid hydrogen storage systems

Quantitative Risk Assessment (QRA) supports the development of risk-informed safety codes and standards which are employed to enable the safe deployment of hydrogen technologies essential to decarbonize the transportation sector. System reliability data is a necessary input for rigorous QRA. The lack of reliability data for bulk liquid hydrogen (LH₂) storage systems located on site at fueling stations limits the use of QRAs. In turn, this hinders the ability to develop the necessary safety codes and standards that enable worldwide deployment of these stations. Through a QRA-based analysis of a LH₂ storage system, this work focuses on identifying relevant scenario and probability data currently available and ascertaining future data collection requirements regarding risks specific to liquid hydrogen releases. The work developed consists of the analysis of a general bulk LH₂ storage system design located at a hydrogen fueling station. Failure Mode and Effect Analysis (FMEA) and traditional QRA modeling tools such as Event Sequence Diagrams (ESD) and Fault Tree Analysis (FTA) are employed to identify, rank, and model risk scenarios related to the release of LH₂. Based on this analysis, scenario and reliability data needs to add LH₂-related components to QRA are identified with the purpose of improving the future safety and risk assessment of these systems.     

乌兰矿采矿现状及技术应对方案

本文分析了乌兰矿投产前期采矿现状及存在的主要问题,针对该矿所处蒙古国经济落后、投资风险大的现实状况,为避免生产中断、规避投资风险,早日回收前期投资考虑,采取了双斜坡道开拓、全尾胶结充填、高端壁空场嗣后充填采矿、多中段组合式连续开采等系列技术应对方案。大大降低了一次性投资规模及投资风险,前期投资得以快速回笼的同时,矿山产能也充分释放,确保了矿山的持续稳定,取得了较好的经济和社会效益。为海外地下近地表矿体开采矿山规避投资风险提供了很好的技术方案借鉴。     

Suckling behavior of calves in seasonally calving pasture-based dairy systems,and possible environmental and management factors affecting suckling behaviors

In recent years, interest has been increasing in whether farmed animals are able to live a reasonably natural life, with one particular area of concern being calf-dam separation. The objectives of this study were to monitor the timing and frequency of suckling behavior of calves left on pasture to suckle their dams for up to 24 h (interquartile range 4.0–15.5 h) and to investigate possible risk factors that may contribute to any variability seen. Over 2 yr, a convenience sample of 8 farms (4 in the North Island, 4 in the South Island of New Zealand) were involved in an observational study where cows and calves were observed for 24 h a day over a 2-wk-long period per farm. During the observation period, farmers continued to remove calves at the same frequency they normally did (which ranged from once a day to 4 times a day). Cows (between 2 and 12 yr old) and calves were observed from a scissor lift in or beside the calving paddock. Cows had numbers written on them, and observers used binoculars and spotlights. Observers recorded the length of stage 2 labor, time of birth, standing, and first suckling, number of suckling events, time of calf removal from the dam, temperature where the cows were grazing, and size of the grazing area they were calving in. Dams were body condition scored before calving, and their age was extracted from farm records. A total of 697 calves were observed during the study. A total of 444 of 697 calves [63.7%; 95% confidence interval (CI) = 60.0–67.3%] suckled in the calving paddock (farm range 40.0% to 90.2%). Of the 444 calves that suckled in the calving paddock, 407 (58.4%; 95% CI 54.6–62.1%) suckled within the first 6 h after birth (farm range 33.0% to 83.6%). Individual risk factors associated with the hazard rate ratio (HR) for time to first suckling event were time to standing (calves who took more than 1.3 h to stand had a longer time from birth to first suckle) and age of the dam [compared with calves that were born from dams >7 years of age, calves born to dams that were 2–3 and 4–7 yr of age had a 1.49 (95% CI 1.07–2.06) and 1.19 (95% CI 0.89–1.60) HR, respectively, for time from birth to first suckle in the calving paddock]. Farm risk factors associated with the HR of suckling were frequency of calf collection [calves that were born on farms that collected calves once a day suckled earlier than calves on farms that removed calves more than once a day (HR 1.52; 95% CI 1.25–1.84)] and temperature [a minimum temperature of <10°C within 6 h of a calf being born was associated with a 0.69 (95% CI 0.53–0.89) hazard of suckling in the calving paddock]. We observed very large farm variability that urgently requires further investigation if pasture-based farms are ever to adopt a system where calves remain with their dams for longer than 24 h.     

基于多通道的ICU脑血管疾病死亡风险预测模型

死亡风险预测指根据病人临床体征监测数据来预测未来一段时间的死亡风险。对于ICU病患，通过死亡风险预测可以有针对性地对病人做出临床诊断，以及合理安排有限的医疗资源。基于临床使用的MEWS和Glasgow昏迷评分量表，针对ICU病人临床监测的17项生理参数，提出一种基于多通道的ICU脑血管疾病死亡风险预测模型。引入多通道概念应用于BiLSTM模型，用于突出每个生理参数对死亡风险预测的作用。采用Attention机制用于提高模型预测精度。实验数据来自MIMIC [Ⅲ]数据库，从中提取3?080位脑血管疾病患者的16?260条记录用于此次研究，除了六组超参数实验之外，将所提模型与LSTM、Multichannel-BiLSTM、逻辑回归（logistic regression）和支持向量机（support vector machine, SVM）四种模型进行了对比分析，准确率Accuracy、灵敏度Sensitive、特异性Specificity、AUC-ROC和AUC-PRC作为评价指标，实验结果表明，所提模型性能优于其他模型，AUC值达到94.3%。