Improvement of region-based motion estimation by considering uncovered regions

In natural video sequences, object movement causes regions to be covered or uncovered. Conventional algorithms for region-based motion estimation do not take uncovered regions into full account. Uncovered regions seriously decrease the accuracy of motion estimation. This paper presents an algorithm for increasing the motion estimation accuracy. This algorithm detects uncovered regions and uses them to improve image segmentation and motion estimation. Experimental results show that the presented algorithm is effective in reducing the displaced frame difference, without introducing any extra information for coding applications.     

Methodologies used for evaluation of video tools and algorithms in MPEG-4

MPEG-4 issued two calls for proposals requesting submission of algorithms and tools relevant to standardization of MPEG-4. This paper reports on the evaluation of tools submitted for evaluation in November 1995 and January 1996. Complete video coding schemes submitted in January 1996 are also covered. The goal of the evaluation was to cluster the tools according to the technical areas they address, to evaluate them according to the issues relevant to the standardization process, and finally to suggest areas of core experiments to improve a video verification model (VM) as soon as the VM becomes available. Altogether, MPEG evaluated 87 tools and 19 complete coding algorithms, most of them highlighted in this paper. During the evaluation, 19 areas for core experiments were identified. Each core experiment is targeted at different functionalities like compression efficiency, content-based coding, error resilience, scalability. This definition of core experiments caused close collaboration and supported mutual fertilization between organizations working on similar tools, which allowed the VM to progress much faster than expected.     

基于GEA运动估计算法的低功耗容错VLSI结构

提出了一种能量高效的运动估计结构.该结构基于GEA (Global Elimination Algorithm)算法,在并行GEA结构的基础上,对关键的绝对差和模块应用差错复原机制,以对抗在工艺参数波动和/或工作电压超比例缩小(VOS)时可能产生的逻辑级时序错误.采用一个亚采样电路ISR-SSAD,将VOS技术和算法级容噪设计集成到绝对差和模块中,实现了差错检测和纠正,与原GEA结构相比,具有更低的功耗.计算结果表明,整个运动估计模块的功率节省可达16%.     

适于运动估值VLSI实现的改进的树结构

文中给出了一种用于实现多分辨率运动估算算法后阶段任务的改进的树结构。在一个简单的RISC类型核控制下,它能够完成整个运动估值过程中除粗分辨率精度运动矢量搜索之外的所有后阶段子任务。包括运动矢量优化（搜索）在内的多任务是通过二叉树最底层叶节点上的多功能处理单元和可以拆分成子树的加法树来实现的。此外,运算单元寄存器堆的设计使能在二维方向上复用图像数据,完全避免了同一类数据从存储器中重复读取,从而实现了最小的存储器访问带宽,并有助于减小存储功耗。     

Remarks on the differential method for the estimation of movement in television images

In this paper a physical interpretation is given of a method for the estimation of movements in television images. The method, already presented by the author in other papers, is based on a linear regression of the image derivatives, both spatial and temporal. It is shown that a centered finite differences approximation of the image differentials is mandatory to obtain good performances. Experimental results are issued to support theoretical conclusions.     

On the strategic origin of Brownian motion in finance

This paper is concerned with the strategic use of a private information on the stock market. A repeated auction model is used to analyze the evolution of the price system on a market with asymmetric information.  The model turns out to be a zero-sum repeated game with one-sided information, as introduced by Aumann and Maschler.  The stochastic evolution of the price system can be explicitly computed in the n times repeated case. As n grows to ∞, this process tends to a continuous time martingale related to a Brownian Motion.  This paper provides in this way an endogenous justification for the appearance of Brownian Motion in Finance theory. Received: February 2002     

基于双目视觉模型的运动参数测量

通过建立双目测量的数学模型,由标志点在双摄像机成像平面上的坐标求得运动物体的运动参数。在不考虑摄像机内外参数标定误差的情况下,研究了标志点像元的测量精度对运动参数测量的影响。通过分析图像分割求取标志点形心引起的误差,利用窗口中心迭代算法提高标志点形心测量的精度,从而提高了运动参数的测量精度。由空间飞行器自动交会的仿真试验验证了算法的有效性。     

Content adaptive video denoising based on human visual perception

In this paper, we propose content adaptive denoising in highly corrupted videos based on human visual perception. We introduce the human visual perception in video denoising to achieve good performance. In general, smooth regions corrupted by noise are much more annoying to human observers than complex regions. Moreover, human eyes are more interested in complex regions with image details and more sensitive to luminance than chrominance. Based on the human visual perception, we perform perceptual video denoising to effectively preserve image details and remove annoying noise. To successfully remove noise and recover the image details, we extend nonlocal mean filtering to the spatiotemporal domain. With the guidance of content adaptive segmentation and motion detection, we conduct content adaptive filtering in the YUV color space to consider context in images and obtain perceptually pleasant results. Extensive experiments on various video sequences demonstrate that the proposed method reconstructs natural-looking results even in highly corrupted images and achieves good performance in terms of both visual quality and quantitative measures.     

基于仿射模型的重叠块运动估计补偿方法

本文对 Ｈ．２６３中基于平移模型的重叠块运动估计补偿方法进行改进,提出了一种基于仿射模型的重叠块运动补偿方法,该方法对复杂运动图像质量有明显改善。     

基于H.263的实时视频编码技术研究

为了提高H．263编码技术现有算法的执行效率，本文提出了一种基于时域预测的预判零方法，该方法在保证一定图像质量的条件下具有较高的判别效率，能节约大量的编码时间：并在此基础上结合“钻石”运动搜索算法、MMX优化技术，实现了利用软件进行实时视频编码。