Learning Image-Text Associations

Web information fusion can be defined as the problem of collating and tracking information related to specific topics on the World Wide Web. Whereas most existing work on Web information fusion has focused on text-based multidocument summarization, this paper concerns the topic of image and text association, a cornerstone of cross-media Web information fusion. Specifically, we present two learning methods for discovering the underlying associations between images and texts based on small training data sets. The first method based on vague transformation measures the information similarity between the visual features and the textual features through a set of predefined domain-specific information categories. Another method uses a neural network to learn direct mapping between the visual and textual features by automatically and incrementally summarizing the associated features into a set of information templates. Despite their distinct approaches, our experimental results on a terrorist domain document set show that both methods are capable of learning associations between images and texts from a small training data set.     

基于光纤环网的隧道火灾监控系统研究

文章对基于光纤环网的隧道群火灾报警系统进行研究,分析了其优缺点,提出了解决办法,并完成改造,应用于实际.     

Layered image coding using the DCT pyramid

A block-based subband image coder that exploits the ability to perform decimation in the discrete cosine transform (DCT) domain to effect a pyramidal data structure is described. The proposed "DCT pyramid" has a distinct feature of improved image rendition properties without the associated blocking artifacts at low bit-rates.     

DNA-Directed alkylating agents. 7. Synthesis, DNA interaction, and antitumor activity of bis(hydroxymethyl)- and bis(carbamate)-substituted pyrrolizines and imidazoles

A series of bis(hydroxymethyl)-substituted imidazoles, thioimidazoles, and pyrrolizines and related bis(carbamates), linked to either 9-anilinoacridine (intercalating) or 4-(4-quinolinylamino)benzamide (minor groove binding) carriers, were synthesized and evaluated for sequence-specific DNA alkylation and cytotoxicity. The imidazole and thioimidazole analogues were prepared by initial synthesis of [(4-aminophenyl)alkyl]imidazole-, thioimidazole-, or pyrrolizine dicarboxylates, coupling of these with the desired carrier, and reduction to give the required bis(hydroxymethyl) alkylating moiety. The pyrrolizines were the most reactive alkylators, followed by the thioimidazoles, while the imidazoles were unreactive. The pyrrolizines and some of the thioimidazoles cross-linked DNA, as measured by agarose gel electrophoresis. Strand cleavage assays showed that none of the compounds reacted at purine N7 or N3 sites in the gpt region of the plasmid gpt2Eco, but the polymerase stop assay showed patterns of G-alkylation in C-rich regions. The corresponding thioimidazole bis(carbamates) were more selective than the bis(hydroxymethyl) pyrrolizines, with high-intensity bands at 5'-NCCN, 5'-NGCN and 5'-NCGN sequences in the PCR stopping assay ( indicates block sites). The data suggest that these targeted compounds, like the known thioimidazole bis(carbamate) carmethizole, alkylate exclusively at guanine residues via the 2-amino group, with little or no alkylation at N3 and N7 guanine or adenine sites. The cytotoxicities of the compounds correlated broadly with their reactivities, with the bis(hydroxymethyl)imidazoles being the least cytotoxic (IC50s >1 microM; P388 leukemia) and with the intercalator-linked analogues being more cytotoxic than the corresponding minor-groove-targeted ones. This was true also for the more reactive thioimidazole bis(carbamates) (IC50s 0.8 and 11 microM, respectively), but both were more active than the analogous "untargeted" carmethizole (IC50 20 microM). The bis(hydroxymethyl)pyrrolizine analogues were the most cytotoxic, with IC50s as low as 0.03 microM.     

基于多目视觉的变形钢板三维测量方法

提出了一种使用4个相机测量变形钢板三维形状的方法;在没有纹理的钢板上画一些点,使用4个相机同时拍摄钢板;利用边缘检测算法找出钢板的边缘,利用特征点提取算法提取出4个图像中钢板上的特征点,并找到4个图像上精确的特征点匹配关系,然后测量出这些特征点的三维点;最后,以这些三维点为种子点向外区域增长,从而得到钢板的三维形状;实验对变形钢板进行三维测量,得到了钢板的精确的三维形状,结果验证了该方法的有效性和可行性。     

On pebble automata for data languages with decidable emptiness problem

In this paper we study a subclass of pebble automata (PA) for data languages for which the emptiness problem is decidable. Namely, we show that the emptiness problem for weak 2-pebble automata is decidable, while the same problem for weak 3-pebble automata is undecidable. We also introduce the so-called top view weak PA. Roughly speaking, top view weak PA are weak PA where the equality test is performed only between the data values seen by the two most recently placed pebbles. The emptiness problem for this model is still decidable. It is also robust: alternating, non-deterministic and deterministic top view weak PA have the same recognition power; and are strong enough to accept all data languages expressible in Linear Temporal Logic with the future-time operators, augmented with one register freeze quantifier.     

Performance Analysis of Gradient Neural Network Exploited for Online Time-Varying Matrix Inversion

This technical note presents theoretical analysis and simulation results on the performance of a classic gradient neural network (GNN), which was designed originally for constant matrix inversion but is now exploited for time-varying matrix inversion. Compared to the constant matrix-inversion case, the gradient neural network inverting a time-varying matrix could only approximately approach its time-varying theoretical inverse, instead of converging exactly. In other words, the steady-state error between the GNN solution and the theoretical/exact inverse does not vanish to zero. In this technical note, the upper bound of such an error is estimated firstly. The global exponential convergence rate is then analyzed for such a Hopfield-type neural network when approaching the bound error. Computer-simulation results finally substantiate the performance analysis of this gradient neural network exploited to invert online time-varying matrices.