Robust image retrieval by cascading a deep quality assessment network

The performance of computer vision algorithms can severely degrade in the presence of a variety of distortions. While image enhancement algorithms have evolved to optimize image quality as measured according to human visual perception, their relevance in maximizing the success of computer vision algorithms operating on the enhanced image has been much less investigated. We consider the problem of image enhancement to combat Gaussian noise and low resolution with respect to the specific application of image retrieval from a dataset. We define the notion of image quality as determined by the success of image retrieval and design a deep convolutional neural network (CNN) to predict this quality. This network is then cascaded with a deep CNN designed for image denoising or super resolution, allowing for optimization of the enhancement CNN to maximize retrieval performance. This framework allows us to couple enhancement to the retrieval problem. We also consider the problem of adapting image features for robust retrieval performance in the presence of distortions. We show through experiments on distorted images of the Oxford and Paris buildings datasets that our algorithms yield improved mean average precision when compared to using enhancement methods that are oblivious to the task of image retrieval. ¹     

简析变压器油的生产工艺及质量

本文简要介绍了变压器油的运用价值,同时分析如何实现对变压器油生产工艺的优化,并从物理性能以及化学性能两个方面分析生产工艺改良后变压器油整体质量,以期为变压器油生产工作提供一定的参考与帮助。     

Social impact assessment of investment activities in the China–Pakistan economic corridor

In order to analyze potential social risks, and achieve smooth implementation of investment activities in the China–Pakistan economic corridor (CPEC), we attempt to establish social impact and risk indicators. Both the impacts on the projects and the impacts of the projects are used to develop risk indicators in accordance with the literature review, reality and international standard principles. Using an analytic hierarchy process (AHP) and average value to analyze questionnaires of both Chinese and Pakistani respondents, we identify the importance of various risk indicators. The primary social risks of the western high mountain region include tribal obstacles and religious extremism. The social risks to the southeast coastal plains include preserving the historical and cultural heritage of the area and international protection of national parks. The social risks in north Kashmir include disputes, extremist threats, religious and cultural differences, and the protection of natural reserves. The social risks of the Xinjiang region primarily involve social issues faced by ethnic minorities and environmental pressures.     

Multi-hazard loss analysis of tall buildings under wind and seismic loads

This paper presents a framework for life-cycle loss estimation for non-structural damage in tall buildings under wind and seismic loads. Life-cycle cost analysis is a useful design tool for decision- makers, aimed at predicting monetary losses over the lifetime of a structure, accounting for uncertainties involved in the problem definition. For tall buildings, sensitive to dynamic excitations like earthquake and wind, it can be particularly suitable to base design decisions not only on initial cost and performance but also on future repair expenses. The proposed approach harmonises the procedures for intervention costs evaluation of structures subjected to multiple-hazards, taking into account the peculiar differences of wind and earthquake, in terms of load characterisation, type and evolution of damage. Relative effect of the two hazards on damage to drift- and acceleration-sensitive non-structural elements are examined. The influence of uncertainty in structural damping is also taken into account. It is shown that, although it is commonly believed that the design of a given structure is usually dominated by either winds or earthquakes, when LCC-based design is performed, both winds and earthquakes may be important.     

Occurrence and risk assessment of mycotoxins,acrylamide, and furan in Latvian beer

This work reports data on the occurrence of nine mycotoxins and two food processing contaminants – acrylamide and furan – in a total of 100 beers produced in Latvia. Mycotoxins were detected by high-performance liquid chromatography (HPLC) coupled with time-of-flight mass spectrometry, acrylamide by HPLC coupled with quadrupole-Orbitrap mass spectrometry, and furan by headspace gas chromatography-mass spectrometry. The most frequently occurring mycotoxins were HT-2 and deoxynivalenol (DON), which were detected in 52% and 51% of the analysed samples. The highest content was observed for DON, reaching the maximum of 248 µg kg^?1. Furan was ubiquitous, and 74% of the samples contained acrylamide. In terms of the estimated exposure, the biggest potential risk was identified for HT-2 representing more than 11% of tolerable weekly intake. The margin of exposure approach indicated the exposure to furan through beer as significant, this parameter being close to the critical limit.     

煤矿薄煤层定向钻孔水害探查技术研究

从保德煤矿水文地质条件出发，研究了矿区奥灰水突水危险性，计算了8号煤层底板安全隔水层厚度和突水系数。采用薄煤层定向钻探工艺对保德煤矿81312工作面胶带运输巷道和81313工作面辅助运输巷道底板以下岩层及10号煤层进行水害及隐伏构造探查，探明了8号煤层底板岩层地质构造及富水性，形成了区域薄煤层定向钻孔水害探查方法。研究表明，利用定向钻孔的一孔多用功能超前预抽10号煤层瓦斯，可为矿区安全开采提供保障。     

Quality-by-design approach for the development of telmisartan potassium tablets

A quality-by-design approach was adopted to develop telmisartan potassium (TP) tablets, which were bioequivalent with the commercially available Micardis^® (telmisartan free base) tablets. The dissolution pattern and impurity profile of TP tablets differed from those of Micardis^® tablets because telmisartan free base is poorly soluble in water. After identifying the quality target product profile and critical quality attributes (CQAs), drug dissolution, and impurities were predicted to be risky CQAs. To determine the exact range and cause of risks, we used the risk assessment (RA) tools, preliminary hazard analysis and failure mode and effect analysis to determine the parameters affecting drug dissolution, impurities, and formulation. The range of the design space was optimized using the face-centered central composite design among the design of experiment (DOE) methods. The binder, disintegrant, and kneading time in the wet granulation were identified as X values affecting Y values (disintegration, hardness, friability, dissolution, and impurities). After determining the design space with the desired Y values, the TP tablets were formulated and their dissolution pattern was compared with that of the reference tablet. The selected TP tablet formulated using design space showed a similar dissolution to that of Micardis^® tablets at pH 7.5. The QbD approach TP tablet was bioequivalent to Micardis^® tablets in beagle dogs.     

Nonzero repair times dependent on the failure hazard

It is common in the literature on the reliability and maintenance of repairable systems to model the repair times as instantaneous. However, this is an unreasonable assumption for some complex systems, especially those requiring a high level of reliability, and such systems may spend a significant proportion of their lifetimes under maintenance and repair. We model the ageing of such a system with alternating stochastic processes. Operational times are generated at random and may have an increasing failure rate. Repair times are generated from a random process where the repair time is related to the hazard rate at failure. This yields lengthened repair times at late stages in a system subject to an increasing failure hazard rate but also accommodates long repair times at young ages in systems with a bathtub-shaped hazard rate function. We derive analytic results for a set of special cases of the model, show how simulation and inference can be carried out, and apply our method to real data from a large car manufacturer.     

No-reference quality assessment of dynamic sports videos based on a spatiotemporal motion model

This paper proposes an approach to improve the performance of no-reference video quality assessment for sports videos with dynamic motion scenes using an efficient spatiotemporal model. In the proposed method, we divide the video sequences into video blocks and apply a 3D shearlet transform that can efficiently extract primary spatiotemporal features to capture dynamic natural motion scene statistics from the incoming video blocks. The concatenation of a deep residual bidirectional gated recurrent neural network and logistic regression is used to learn the spatiotemporal correlation more robustly and predict the perceptual quality score. In addition, conditional video block-wise constraints are incorporated into the objective function to improve quality estimation performance for the entire video. The experimental results show that the proposed method extracts spatiotemporal motion information more effectively and predicts the video quality with higher accuracy than the conventional no-reference video quality assessment methods.