2Mb／s电视会议系统

本文介绍了一个自行研制的电视会议系统。该系统由三个终端和一个多点控制器构成。它已在沈阳、长春和哈尔滨三点之间通过数字微波信道实现了会议电视通信，信道速率为２Ｍｂ／ｓ。实践表明，该系统有良好的图象质量和声音质量。本文介绍了系统构成、系统的主要性能、工作原理及设计原则。对Ｈ．２６１图象压缩编码与Ｇ．７２２声音压缩编码部分作了较详细的介绍。     

基于小波变换的Dirac视频编解码器分析

Dirac视频编解码算法是BBC提出的一套方案,同时给出比特流语法及软件实现的开源代码且无专利。本文对其中关键技术进行分析:采用块交叠运动补偿和小波变换进行自适应算术编码的混合编解码方案。使用msu视频质量测试软件对Dirac和Snow及X264在低速运动、快速运动、混合运动情况下的图像质量分别进行了比较。实验结果表明,Dirac的图像质量最稳定,但存在赝像;X264在对象快速运动时质量较差;Snow质量居中。最后分析了Dirac目前存在的问题,进一步研究方向及应用前景。     

基于嵌入式的PSO-BP神经网络码率控制算法研究

在多变的4G移动网络信道中,针对嵌入式移动端视频传输时出现的网络拥塞现象,利用神经网络与码率控制算法相结合的方法,减轻或预防网络拥塞的发生。接收端产生RTCP包并解析出最能直接反映信道情况的延时、丢包和抖动3个信道参数,同时将其作为输入参数反馈给神经网络,将想要预测的码率值作为输出参数,X264编码器根据最佳预测码率更新编码速率。对BP神经网络和PSO-BP神经网络算法编码速率的预测结果进行了比较,说明优化过的码率控制算法的预测误差更小。为了贴近实际,NS2仿真通过添加不同干扰流的实验及分析得到预测性能和结果,该方法可以将反映视频质量的参数PSNR值控制在32~37之间,码率的预测误差控制在±60 kb/s左右。实验结果表明,该算法在拥塞情况下能够有效调整码率,实现流畅、清晰的视频传输。     

Handoff resource prediction in multimedia wireless networks

This paper presents a local-based approach for predicting handoff resources in multimedia wireless networks, where the underlying prediction mechanism is the recursive least-squares (RLS) algorithm. Performance of the proposed handoff resource prediction scheme is evaluated for a multimedia wireless network characterized by various handoff traffic conditions, different bandwidth requirements, and general call and channel holding time distributions. The accuracy of handoff resource prediction is measured in terms of a mean-square-error metric. Results from performance evaluation show that the proposed RLS-based scheme has better resource prediction accuracy than Wiener-filter and autoregressive (AR) processes. It is shown that the proposed RLS-based handoff resource prediction scheme achieves capacity gains of up to 60% and 23% compared to the Wiener- and AR-based schemes, respectively.     

电流耦合型人体通信准静态建模与收发器设计

建立电流耦合型人体通信准静态场的电磁模型,为收发器设计提供理论指导。采用有限元方法建立包含电容效应的人体上臂模型,分析人体内部耦合电流的传输特性。以FPGA为平台设计出FSK调制器、全数字锁相环位同步电路和信号解调器,并辅以发送端低通滤波、信号保持和接收端前处理电路,构成电流耦合型人体通信收发器。模型的传输增益特性具有高通特性,在100 kHz以下与实验数据具有较好的一致性。收发器可以5 Kb/s速率实现数字基带信号的可靠通信。建模与设计工作为电流耦合型人体通信的实现提供理论与实践依据。     

Inner code assisted parallel concatenated single parity check codes

This paper proposes a novel coding scheme to improve the performance of multidimensional parallel concatenated single parity check (PCSPC) codes. The high error floor of PCSPC codes prevents clear turbo cliff to be seen in the bit‐error rate (BER) performance. Based on the product accumulate type‐I (PA‐I) coding scheme, which adds serially an accumulator to the existing PCSPC code structure, our coding scheme considers a wide range of rate‐1 recursive systematic convolutional (RSC) codes replacing the accumulator. The convergence behavior of the proposed iterative decoding is monitored and analyzed using an extrinsic information transfer (EXIT) chart. It has been revealed from the EXIT chart analysis that the proposed coding scheme has a lower convergence threshold than the PA‐I coding scheme, especially when using a lower code rate PCSPC code as the outer code. For the 2D PCSPC(3,2) code with the code rate 0.5, the proposed coding scheme has a convergence threshold of 0.47 dB, whereas the convergence threshold of the PA‐I coding scheme is 0.81 dB. The convergence threshold of the proposed coding scheme is closer to theoretical limit (0.28 dB away from the theoretical limit) than that of the PA‐I coding scheme (0.62 dB away). However, for PCSPC codes with a higher code rate, the performance improvement of the proposed coding scheme with respect to the PA‐I coding scheme becomes smaller. A series of simulations is carried out to investigate the performance of the proposed coding scheme, and the simulation results show that better performance is achieved by using the proposed coding scheme compared with the PA‐I coding scheme. The simulation results show good consistency with the convergence threshold obtained from the EXIT chart, as the difference is only within 0.34 dB in all the evaluated cases. © 2014 Institute of Electrical Engineers of Japan. Published by John Wiley & Sons, Inc.     

An efficient inter prediction mode selection scheme for advanced video coding based on motion homogeneity and residual complexity

Modern video coding standards such as High Efficiency Video Coding (HEVC) and H.264/MPEG‐4 advanced video coding (AVC) supersede the previous coding standards because of their improved coding efficiency. These standards adopt variable block sizes in frame coding ranging from 4 × 4 to 16 × 16 and 4 × 4 to 64 × 64 for H.264/AVC and HEVC, respectively. The use of variable block sizes for inter prediction provides a significant coding gain compared to coding a macroblock (MB) using regular block size. However, this new feature greatly increases the computational complexity of the encoder when brute‐force rate distortion optimization (RDO) algorithm is used for coding parameter selection. This paper proposes an efficient inter prediction mode selection scheme based on motion homogeneity and residual complexity measures of an MB to speed up the encoding process. The motion homogeneity is assessed through the normalized motion vector (MV) field, and residual complexity is evaluated by the sum of absolute difference (SAD). To acquire the MVs and SADs, motion estimation at 8 × 8 block size is performed using a lightweight recursive motion estimator in which the vector field tends toward true object motion. Based on motion homogeneity and residual complexity of an MB, only a small number of inter prediction modes are selected for the RDO process. The experimental results for H.264/AVC show that the proposed scheme reduces the encoding time by 64% on average without any significant degradation of coding efficiency. © 2016 Institute of Electrical Engineers of Japan. Published by John Wiley & Sons, Inc.     

H.264帧内预测模式判定快速算法研究

帧内预测是 H.264编码中用来提升帧内编码压缩效率的编码工具,它通过计算和比较多种帧内预测模式下的宏块率失真代价,并选择最优的帧内预测模式来充分消除图像信号的空间冗余,提升视频信号帧内编码的压缩效率,但多种帧内预测模式下的宏块率失真代价计算使帧内编码运算量成倍增长,阻碍了其实际应用.为此,提出一种帧内预测模式判定快速算法,以提升 H.264帧内编码的计算效率.该算法包括3个部分:1)加快4×4帧内预测模式判定的两级预测模式判定快速算法;2)加快宏块帧内预测模式判定的亮度分量帧内预测分区类型提前判定算法;3)融合上述2种算法的宏块帧内预测模式快速判定流程.测试结果显示,与H.264标准参考编解码器 JM10.1相比,所提算法编码时间节省52%左右,峰值信噪比仅下降0.03 dB,码率增长约为2.5%.由此证明,所提算法能在保持视频压缩效率基本不变的前提下,有效提升视频编码的运算效率     

CNN‐based architecture for real‐time object‐oriented video coding applications

This paper presents a new CNN‐based architecture for real‐time video coding applications. The proposed approach, by exploiting object‐oriented CNN algorithms and MPEG encoding capabilities, enables low bit‐rate encoder/decoder to be designed. Simulation results using Claire video sequence show the effectiveness of the proposed scheme. Copyright © 2005 John Wiley & Sons, Ltd.     

基于T-S模型的电力变压器顶层油温预测研究

为准确估算电力变压器绕组热点温度，对变压器安全运行和寿命评估提供辅助依据，建立了一个基于Takagi- Sugeno(T-S))的变压器顶层油温预测模型。模型的前件参数由模糊C均值聚类算法确定，后件参数由加权最小二乘法离线辨识，并用现场实测数据对模型后件参数进行在线调整。通过对实测数据的仿真实验表明，该模型以简单的模糊规则实现了变压器顶层油温的预测，且模型的预测精度优于IEEE推荐的变压器顶层油温经验模型，从而提高其绕组热点温度计算的精度。     

移动视频业务的EQoS方案及跨层资源管理模型

在分析4G网络的链路特性、视频业务对网路资源的要求和3GPP移动网路QoS方案的基础上,针对移动视频业务在未来ICT网络中的应用特点,提出了移动视频业务扩展QoS（EQoS）方案和基于EQoS方案的跨层资源管理模型。前者引入了移动视频业务细化衡量标准,有利于移动视频业务质量指标的量化;后者在保持原有网络分层模型的基础上,通过EQoS模块实现了跨层无线资源管理。     

流媒体技术及应用

流媒体技术是一种无需等待音／视频信息全部下载,可以随时连续传递和播放信息的传输技术。文中简要介绍了流媒体的基本技术原理及其压缩编码技术,最主要的几种网络传输协议和流媒体的主要传播方式,分析了传统视频告警系统的不足,提出了基于流媒体技术,由流媒体编码器、服务器和播放器的数字视频告警系统的实施方案。     

Integrated rate, power and cell control in cellular CDMA systems: an interference-balancing approach

In this paper, a novel combined rate, power and cell control scheme that minimizes transmit bit energy while gracefully implementing the congestion control is proposed and studied. The reverse link in a code division multiple access (CDMA) system with a high degree of traffic fluctuation temporally as well as spatially is considered. The proposed radio resource management scheme attempts to reduce the interference power variation among base station antennas that can exist in a real radio network. Transmission rates of those users in the congested (non-congested) cells are decreased (increased); hence, the proposed scheme is appropriate for delay-insensitive variable-bit-rate applications. Minimization of average transmit bit energy is achieved, subject to maintenance of individual target Eb/I0 for each user. Two algorithms, one directly minimizing the transmit bit energy, and the other doing so indirectly using the measured pilot power, are presented. It can be shown that both algorithms select the optimal base station if forward- and reverse-link gains are reciprocal; however, the latter gain is decentralized and uses only the local measurements and is amenable for practical implementation.     

Practical low-cost visual communication using binary images for deaf sign language.

Deaf sign language transmitted by video requires a temporal resolution of 8 to 10 frames/s for effective communication. Conventional videoconferencing applications, when operated over low bandwidth telephone lines, provide very low temporal resolution of pictures, of the order of less than a frame per second, resulting in jerky movement of objects. This paper presents a practical solution for sign language communication, offering adequate temporal resolution of images using moving binary sketches or cartoons, implemented on standard personal computer hardware with low-cost cameras and communicating over telephone lines. To extract cartoon points an efficient feature extraction algorithm adaptive to the global statistics of the image is proposed. To improve the subjective quality of the binary images, irreversible preprocessing techniques, such as isolated point removal and predictive filtering, are used. A simple, efficient and fast recursive temporal prefiltering scheme, using histograms of successive frames, reduces the additive and multiplicative noise from low-cost cameras. An efficient three-dimensional (3-D) compression scheme codes the binary sketches. Subjective tests performed on the system confirm that it can be used for sign language communication over telephone lines.     

AVS软件解码器瓶颈分析与优化

介绍了AVS视频编解码标准的研究意义、关键技术、解码原理,给出了AVS解码器的原理图和软件解码器实现一帧解码的关键流程图。在PC机上采用Intel Vtune工具分析了解码器的耗时情况,接着对解码器的C代码进行了整体性优化,并针对瓶颈模块中的亮度插值函数、反变化和反变化部分提出了一些具体的优化方法。实验结果表明,在保证解码质量的前提下,该优化方案大大提高了解码速度,基本上能达到实时解码的要求。     

基于HTTP的流媒体自适应码率切换算法

基于HTTP的动态流媒体自适应(DASH)系统已经广泛地应用于互联网上的视频传输.针对客户端缓冲区受限和网络环境不稳定的情况,提出了一种新的基于缓冲区的码率控制算法.该算法将缓冲区分成不同等级,并采用滑动窗技术对视频的下载时间进行平滑处理,使缓存尽最大可能地维持在均衡区内.该算法在向上切换码率时采用保守策略,而在向下切换时则利用逻辑斯谛方程来决定码率切换的幅度.此外,该算法采用快速启动策略,能使系统开始时下载码率迅速匹配网络带宽.实验结果表明,算法能在复杂的网络环境中,充分利用带宽和缓冲区,降低码率的切换次数.     

基于数字预失真技术的功放线性化方法研究

针对现代通信技术对功率放大器高线性化的需求,提出了基于数字预失真技术的优化方案,使用记忆多项式功放行为建模逼近真实的非线性特性,并在间接学习结构上采用递归预测误差训练算法迭代预失真权向量,实现对基带信号的校正处理。仿真结果表明该方案对功放非线性起到优化作用,减弱了频谱的泄露。同时,在实测过程中适当引入功率回退,该方案能获得更快的收敛速度和较好的稳定性,对需求高线性度的功放设计有一定的参考价值。     

Resolution of a digital servo control system using single‐bit digital signal processing

The application of single‐bit digital signal processing to mechanical control systems has already been proposed by the authors. Multibit A/D converters have been improved to a high level. But it is difficult to improve high‐resolution A/D by the latest semiconductor technologies. A single‐bit digital signal can be generated by a delta‐sigma modulator. Single‐bit digital signal has a small quantized error at low frequency. Then with this single‐bit digital signal processing, high resolution on controlling such a narrow‐bandwidth mechanical system will be realized. In this paper, resolution of analog, multibit, and single‐bit control systems is estimated via simulation. According to the results of simulations, a single‐bit control system has a higher resolution than a multibit system under conditions of equal bit rates. © 1999 Scripta Technica, Electr Eng Jpn, 128(4): 94–101, 1999     

基于特征相关分析修正的GPSO-LSTM短期负荷预测

针对多因素互影响造成负荷预测精度低的问题,提出一种基于特征相关分析修正与全局粒子群优化(GPSO)的长短期记忆循环神经网络(LSTM)短期负荷预测新方法.该方法首先对负荷相关序列进行探索性数据分析(EDA)及预处理,找寻特征内在机理与相关联系并加以修正,保证输入特征的强相关性和完整性.针对传统前馈神经网络无法处理序列关...     

基于指数率失真模型的增量式码率控制算法

高速增长的视频通信需求促进了视频压缩技术的快速发展。为了进一步提高H.264/AVC的码率控制性能,提出了一种新的码率控制算法。新算法包含一种新型的指数型率失真模型,一种增量式QP计算方法,一种基于权重的帧比特位分配策略和一种线性回归模型参数更新方法。基于指数率失真模型的一阶微分特性和拉格朗日中值定理,提出的增量式QP计算方法成功地解决H.264/AVC中的"QP悖论"。编码实验结果表明,新算法与JM模型中推荐的码率控制算法相比,实际比特率更接近目标比特率,达到了更精确的缓冲控制,明显地提高了平均解码视觉质量QCIF为0.23dB,CIF为0.75dB,720p为0.12dB。