Region-based surveillance video retrieval with effective object representation

Retrieving the most relevant video frames that contain the object specified in a given query (query-by-region) remains a challenging task. Two common challenges of region-based retrieval approaches are to accurately extract or segment object(s) and select a proper matching strategy. This paper addresses these problems by proposing a retrieval approach that uses a new region-based matching technique equipped with an effective object representation method. In the first stage, the proposed approach selects the most informative instances of each object that appeared in the video by utilizing an adapted clustering algorithm over the extracted features. In the retrieval stage, the new matching technique returns the most relevant sequences of video by mapping a given region with those identified representative instances of objects based on their similarity scores. The proposed approach is evaluated on standard datasets and the results demonstrate a 31% improvement in the retrieval performance compared to other state-of-the-art methods.     

New Insight into Procedure of Interface Electron Transfer through Cascade System with Enhanced Photocatalytic Activity

Recombination of photogenerated electron–hole pairs is extremely limited in the practical application of photocatalysis toward solving the energy crisis and environmental pollution. A rational design of the cascade system (i.e., rGO/Bi₂WO₆/Au, and ternary composites) with highly efficient charge carrier separation is successfully constructed. As expected, the integrated system (rGO/Bi₂WO₆/Au) shows enhanced photocatalytic activity compared to bare Bi₂WO₆ and other binary composites, and it is proved in multiple electron transfer (MET) behavior, namely a cooperative electron transfer (ET) cascade effect. Simultaneously, UV–vis/scanning electrochemical microscopy is used to directly identify MET kinetic information through an in situ probe scanning technique, where the “fast” and “slow” heterogeneous ET rate constants (K_eff) of corresponding photocatalysts on the different interfaces are found, which further reveals that the MET behavior is the prime source for enhanced photocatalytic activity. This work not only offers a new insight to study catalytic performance during photocatalysis and electrocatalysis systems, but also opens up a new avenue to design highly efficient catalysts in photocatalytic CO₂ conversion to useful chemicals and photovoltaic devices.     

面向大规模数据集的近邻传播聚类

近邻传播聚类在计算过程中需构建相似度矩阵,该矩阵的规模随样本数急剧增长,限制了算法在大规模数据集上的直接应用。为此,提出一种改进的近邻传播聚类算法,利用数据点的局部分布,借鉴半监督聚类的思想构造稀疏化的相似度矩阵,并对聚类结果中的簇代表点再次或多次聚类,直至得到合适的簇划分。实验结果表明,该算法在处理能力和运算速度上优于原算法。     

An immune inspired co-evolutionary affinity network for prefetching of distributed object

In distributed object computing (DOC) containers, cache strategy as a passive approach improves the performance by caching the recently accessed objects. By the advent of large-scale enterprise applications of DOC, caching methods fail to keep pace with the increasing importance of performance and the increasing scale of DOC system. Prefetching is an effective approach to improve the performance. It generates and stores pages or objects in caches in advance, by predicting the requests. In the current DOC container, prefetching strategy is not supported and seldom studied in the literature. The immune system with faster secondary response and its affinity network inspire applying the immune mechanisms to build holistic model of DOC for performance improvement. In this study, a co-evolutionary affinity network (CEA-Net) is proposed for prefetching distributed objects. In CEA-Net, objects are antibodies and computation tasks are antigens. Invocation relations among object classes are modeled by the immune network of antibody genotypes. Multiple affinity measures among antibody, antigen, genotypes of antibody and antigen, genotype set and antibody population are defined to model the complex relations among distributed objects. Especially in the antibody population, immune principles including clonal proliferation, immune memory, immune toleration and elimination are designed to add evolutionary features to the antibody population. Based on CEA-Net, the prefetching architecture for DOC is built including 5 main procedures, Network Abstractor, Access Recorder, Object Factory, Cache Engine and Prefetch Engine. Finally, the experimental study shows the promising access performance and the evolutionary features of CEA-Net. CEA-Net is instructive for the future design of high performance DOC containers, such as WebSphere Application Server and BEA WebLogic Application Server.     

2HL055恒流二极管参数测试

2HL055是贵州博越公司生产的恒流二极管,该类器件恒定电流大,极限电压低,可以直接驱动负载,但器件形成实验室成品后,没有进行过参数测试,其基本参数范围不能确定。文章拟通过试验,进一步测试2HL055器件的基本参数,并确定部分参数的测试方法。     

基于转台的线阵激光雷达地面检校方法研究

介绍了一种基于转台的线阵激光雷达地面检校方法,该方法通过布设物方检校控制场,使用指示激光器建立激光测距值与物方控制点之间的对应关系并引入激光测距的加/乘常数与外方位元素一并作为检校参数,通过坐标转换列出误差方程式;最后使用最小二乘平差解算所有的检校参数。文中给出了实测数据的实验结果分析及精度评定,验证了本文方法的可行性。     

Multiphase Image Segmentation and Modulation Recovery Based on Shape and Topological Sensitivity

Topological sensitivity analysis is performed for the piecewise constant Mumford-Shah functional. Topological and shape derivatives are combined in order to derive an algorithm for image segmentation with fully automatized initialization. Segmentation of 2D and 3D data is presented. Further, a generalized Mumford-Shah functional is proposed and numerically investigated for the segmentation of images modulated due to, e.g., coil sensitivities.

基于ARM9的恒压恒速双缸泵控制系统

介绍基于ARM9的嵌入式恒压恒速双缸泵控制系统。以S3C2410嵌入式处理器为核心,扩展其外部硬件设备包括高精度AD, DA转换器和LCD等,以Windows CE为操作系统,应用PB和EVC++4.0开发应用软件。该控制系统成本低,界面友好,可实现对恒压恒速双缸泵的高精度高稳定性控制。     

一种脆弱的图像自适应置乱算法

证明在离散余弦变换和小波变换中,图像经空域置乱前后在频域内不变的系数规律,根据构造的复合混沌迭代动力系统,提出2个利用不变频域变换系数规律的构造函数,生成复合混沌迭代动力系统需要的初始值,给出脆弱的图像自适应置乱算法以及相似度和抗攻击半径的置乱算法评价概念。实验表明,算法无需任何附加信息即可提取出原始图像,对各种攻击十分敏感,平均抗攻击半径高于一般算法,适用于法庭、医学等要求极其严格的场合。