Surveillance video synopsis in the compressed domain for fast video browsing

The traditional pixel-domain based video analysis methods have taken dominated places for long. However, due to the rapidly increasing volume and resolution of surveillance video, the desirable fast and scalable browsing encounters significant challenges in terms of efficiency and flexibility. Under this circumstance, operating surveillance video in compressed domain has aroused great concern in academy and industry. In order to perform the intelligent video analysis task on the premise of preserving accuracy and controlling complexity, this paper presents a compressed-domain approach for massive surveillance video synopsis generation, labeling and browsing. The main work and achievements include: (1) a compressed-domain scheme is established to condense the compressed surveillance video and record the synopsis results; (2) a background modeling method via the Motion Vector based Local Binary Pattern (MVLBP) is introduced to extract moving objects in an efficient way; (3) an object flags based synopsis labeling method is proposed to represent the object regions as well as their display modes in a flexible way. Experimental results show that the video analysis system based on this framework can provide not only efficient synopsis generation but also flexible scalable or playback browsing.     

Hierarchically compressed wavelet synopses

The wavelet decomposition is a proven tool for constructing concise synopses of large data sets that can be used to obtain fast approximate answers. Existing research studies focus on selecting an optimal set of wavelet coefficients to store so as to minimize some error metric, without however seeking to reduce the size of the wavelet coefficients themselves. In many real data sets the existence of large spikes in the data values results in many large coefficient values lying on paths of a conceptual tree structure known as the error tree. To exploit this fact, we introduce in this paper a novel compression scheme for wavelet synopses, termed hierarchically compressed wavelet synopses, that fully exploits hierarchical relationships among coefficients in order to reduce their storage. Our proposed compression scheme allows for a larger number of coefficients to be stored for a given space constraint thus resulting in increased accuracy of the produced synopsis. We propose optimal, approximate and greedy algorithms for constructing hierarchically compressed wavelet synopses that minimize the sum squared error while not exceeding a given space budget. Extensive experimental results on both synthetic and real-world data sets validate our novel compression scheme and demonstrate the effectiveness of our algorithms against existing synopsis construction algorithms. This work has been funded by the project PENED 2003. The project is co-financed 75% of public expenditure through EC—European Social Fund, 25% of public expenditure through Ministry of Development—General Secretariat of Research and Technology and through private sector, under measure 8.3 of OPERATIONAL PROGRAMME “COMPETITIVENESS" in the 3rd Community Support Programme.     

基于运动目标轨迹优化的监控视频浓缩方法

视频浓缩是包含原视频有效信息的简短表示,以便于视频的存储、浏览和检索。然而,大部分视频浓缩方法得到的浓缩视频中会丢失少量目标,不能完整表达原始视频的全部内容。本文介绍了一种基于目标轨迹优化的视频浓缩方法。首先使用改进的目标轨迹提取算法提取原视频中目标的 轨迹,然后利用马尔可夫随机场模型和松弛线性规划算法得到每条轨迹的最优时间标签,将其与背景序列和目标轨迹结合生成浓缩视频。实验结果表明,与传统的视频浓缩方法相比,本文方法生成的浓缩视频具有较高的浓缩比,保证了信息的完整性又具有良好的视觉效果。     

基于时空划分的数据流挖掘

基于时空划分的思想,设计概要数据结构的在线生成算法。概要数据结构保存流数据不同时刻的分布状态,以支持离线阶段的分类、聚类和关联规则发现等数据挖掘操作。研究时间粒度、量化向量调整和子区域索引等3项内存需求控制策略,以平衡概要数据结构的内存需求和内外存之间的I/O次数。     

基于数据概要描述的分布式数据流聚类模型与算法

数据流挖掘可有效解决大容量流式数据的知识发现问题,并已得到广泛研究.数据流的一个典型的例子是传感器采集的流式数据.然而,随着传感器网络的应用普及,这些流式数据在很多情况下是分布式采集和管理的,这就必然导致分布式地挖掘数据流的需求.分布式数据流挖掘的最大障碍是由分布式而导致的挖掘质量或者效率问题.为适应分布式数据流的聚类挖掘,探讨了分布式数据流的挖掘模型,并且基于该模型设计了对应的概要数据结构和关键的挖掘算法,给出了算法的理论评估或者实验验证.实验说明,提出的模型和算法可以有效地减少数据通信代价,并且能保证较高的全局模式的聚类质量.     

数据挖掘取样方法研究

取样是一种通用有效的近似技术.在数据挖掘研究中,取样方法可显著减小所处理数据集的规模,使得众多数据挖掘算法得以应用到大规模数据集以及数据流数据上.通过对应用于数据挖掘领域的代表性取样方法的比较研究和分析总结,提出了一个取样算法分类框架.在指出了均匀取样局限性的基础上阐述了某些应用场景中选用偏倚取样方法的必要性,综述了取...     

基于考研大纲的计算机组成原理教学研究

针对日常教学和实验内容与计算机组成原理统考大纲的差异性,提出了基于计算机专业研究生入学考试大纲的教学和实验内容设计,并对目前教学方法提出了几点改进措施,实现了教学效果和考研成绩的双赢。     

一种适用于数据流梗概计算的小波变换算法

针对数据流上近似查询中的梗概计算，提出了一种新的基于最小误差的维压缩小波变换算法（MEDC）．MEDC算法通过映射流数据时间戳，快速无冗余地维护流数据的有序性；基于最小误差，高效压缩小波变换阵列，最大化MEDC算法时间效率及近似查询实时处理能力；引入小波系数与查询准确度之间的数值性关联规则，支持小波系数梗概上的查询多级共享，整体查询执行性能最佳．实验表明，与传统小波变换、直方图和采样等算法相比，MEDC算法在数据流近似查询处理的响应速度、查询结果质量等方面具有更为优越的性能．     

基于新视角合成的视频摘要交互式浏览

现有视频摘要方法大多仅在固定视角下对视频内容进行压缩,限制了视频内容的浏览空间,导致压缩后的视频内容相互碰撞和遮挡.最近的方法改变视频内容的时空关系,但导致摘要结果与原始视频的相似度降低.本文研究基于新视角合成技术的交互式摘要视频浏览算法,通过为视频内容构建新的观察视角,增加视频浏览空间,从本质上减少视频内容碰撞和遮挡的可能性.为完全避免碰撞现象,提出了特定视角下投影遮挡消除技术,仅需较小地改变视频内容的时空关系,保证摘要视频与原始视频内容的一致性.最后,为方便用户交互式浏览摘要视频,构建了观察视角自动生成算法,根据当前摘要内容实时推荐并调整到最优观察视角.实验结果和用户调查表明本文提出的交互式视频摘要算法能有效压缩视频冗余信息,提高了摘要视频准确度、易理解程度.     

基于改进ViBe算法的视频浓缩

针对监控视频在时间上存在冗余的问题,对ViBe(visual background extractor)算法进行改进,解决了ViBe算法存在噪声和易引入鬼影的问题,通过改进后的算法对视频进行背景建模,并对得到的背景掩模提取外轮廓以确定视频帧中是否存在前景对象。将存在前景对象的视频帧写入视频流中,达到视频浓缩的目的。经过试验验证,该方法可以有效地减少视频中的冗余信息,减小视频的体积,视频中的重要信息同时也得到了完整保留,满足实时性要求。