Combined and iterative form of spatial and temporal error concealment for video signals

Video transmission over unreliable channels may suffer from the data loss or corruption. To facilitate the transmission, video content is compressed to save the bandwidth. The compression algorithms remove the redundancies in the video signal, meanwhile increasing the dependencies among symbols in the compressed bit-stream. Thus, when errors occur, they may propagate in both space and time. Among various error control techniques, error concealment (EC) is an effective method that is performed at the decoder to mitigate the influence of errors on the quality of reconstructed images. In this paper, a joint spatial and temporal EC algorithm (JSTEC) is presented. First, several existing temporal EC algorithms are combined in an appropriate order. Then, the spatial correlation in the video is exploited in an iterative form to improve the performance of the temporal EC. Experimental results show that JSTEC performs better both in peak signal-to-noise ratio and subjective quality of images than the existing algorithms.     

基于模糊AHP的TOPSIS评估方法研究

对于模糊环境下的评估,存在固有的不确定性和不精确性信息比较的情况。在AHP的框架下,通过三角模糊数表示元素之间的比较比率,定义了两个模糊数之间的距离,利用模糊综合扩展分析方法构造决策矩阵。根据理想近似方法,通过计算到正理想解和负理想解的距离求出每个方案的贴近系数,得出候选者的评估顺序,并举例说明了应用这个方法的计算过程。     

基于XML建模的研究及其在CAPP系统中的应用

XML是数据访问领域的最新技术,本文分析了基于XML建立数据模型的必然趋势,讲述了基于XML进行模型设计的通用过程。详细分析了CAPP系统中的信息,在此基础上,讲述了基于XML建立CAPP系统信息模型的方法,并且对于零件特征信息进行了建模示例。     

SUPANET的入网控制技术

本文以西南交通大学四川省网络通信技术重点实验室提出的“单物理层用户数据传输平面的体系结构”(SUPA)为背景,通过分析比较现存的各种接纳控制策略和算法,提出了适合SUPA的接纳控制策略和算法。     

实验室服务器规划及其安全策略探讨

阐述了实验室教学管理服务器的规划方法,并总结了服务器系统在安装设置和管理中的一些主要安全策略。     

转炉炼钢脱氧工艺的优化

论述了转炉炼钢脱氧工艺优化的原理,并结合实际对脱氧剂使用工艺进行了探索,成功地在炼钢生产中推广应用.     

元数学对计算机科学的启示

文章提出了元数学对计算机科学的几点启示,这几个方面分别是传统程序设计的概念、机制、要素,数据库技术中的数据模型和函数依赖,面向对象分析和设计方法。文中一方面分析了元数学所提供的理论基础,另一方面说明了在应用方面是如何体现这些理论基础的,并提出了在计算机技术领域中应从数学基础问题、数理逻辑等的相关理论着手寻找突破的思想。     

Quantum Corrected Monte Carlo Simulation of Semiconductor Devices Using the Effective Conduction-Band Edge Method

In recent years much effort has been devoted to the quantum mechanically corrected Monte Carlo simulation of nano-scale semiconductor devices. The effective potential and the Wigner or the Bohm quantum potential are usually chosen for making such quantum correction. In this study we propose another approach based on the Bohm potential by transforming the density gradient to an effective conduction-band edge gradient. Similar to the effective potential method, this method has the advantage of not being affected by density fluctuations. We have applied this method to the simulation of a double-gate nMOSFET. Results are compared to those obtained from the semi-classical Monte Carlo simulation without the quantum correction.     

A Quantum Correction Model for Nanoscale Double-Gate MOS Devices Under Inversion Conditions

A quantum correction model for nanoscale double-gate MOSFETs under inversion conditions is proposed. Based on the solution of Schrödinger-Poisson equations, the developed quantum correction model is optimized with respect to (i) the left and right positions of the charge concentration peak, (ii) the maximum of the charge concentration, (iii) the total inversion charge sheet density, and (iv) the average inversion charge depth, respectively. This model can predict inversion layer electron density for various oxide thicknesses, silicon film thicknesses, and applied voltages. Compared to the Schrödinger-Poisson results, our model prediction is within 3.0% of accuracy. This quantum correction model has continuous derivatives and is therefore amenable to a device simulator.