基于H.264/AVC的压缩域运动对象分割方法

目前大部分压缩域视频对象的分割方法主要面向MPEG系列视频标准,且算法建模复杂。为了解决这一问题,现提出了一种新的基于H.264/AVC的压缩域时空联合运动对象分割（TSMOS）算法。该方法主要利用压缩码流中的DCT系数和运动矢量信息进行对象分割,并首先利用相邻帧DCT系数之差提取运动对象轮廓,同时通过对轮廓进行形态学和抗噪声处理来得到粗糙的运动对象帧差掩码;然后采用运动向量归一化、噪声向量滤除、权值扩展向量中值(WEVM)滤波及前帧分割结果后向投影技术来得到对象的运动掩码;最后通过引入有效机制合并帧差掩码和运动掩码来分割运动对象。实验证明,该算法可取得较好的分割效果。     

一种基于梯度模型的MPEG压缩域的运动对象分割算法

视频数据大都是经过压缩域的形式存储和传输的,且直接在压缩域进行视频对象分割无需运动估计等复杂的计算,速度较快。本文提出了一种基于梯度模型的MPEG压缩域的运动对象分割算法。首先利用DCT(AC[1]和AC[8])系数获得所有物体的边缘,然后综合在累积运动矢量基础上得到的边缘运动信息,从而获得感兴趣运动物体的边缘。仿真实验结果表明,它可以取得满意的分割质量。     

基于H．264压缩域的实时运动对象分割算法

在压缩域内直接分割运动对象对于有实时要求的应用而言是十分必要的，H．264以其优越的压缩效率已经在许多应用中逐渐取代了MPEG-2／4，但有关在H．264压缩域内进行运动对象分割的研究还很少。为此提出了一种从H.264压缩域实时分割运动对象的算法，该算法首先对从H．264视频中提取出的原始运动矢量场进行时域和空域的归一化，接着通过对连续多帧的运动矢量场进行累积来增强显著的运动信息；然后对累积运动矢量场进行全局运动补偿，同时利用快速的统计区域生长算法按照运动相似性将其分割成多个区域；最后利用运动矢量场的方向角直方图来判断出属于运动对象的分割区域，以组成运动对象。通过对多个MPEG-4测试序列的实验结果表明，该方法不仅能够从H．264压缩域中实时地分割出运动对象，且具有良好的分割质量。     

视频监控中一种完整提取运动目标的检测算法

提出一种视频监控中完整、精确提取运动目标前景的检测算法.首先对彩色图像建立混合高斯模型,由背景差分法得到基本准确的前景图像;然后和对称差分法图像综合,得到完整可靠的运动目标图像;再利用亮度信息消除运动目标阴影;最后利用形态学滤波和连通区域面积检测进行后处理.实验结果表明,该算法检测的运动目标前景信息完整准确,对固定场景下的视频监控系统具有一定实用价值.     

Real-time video object segmentation in H.264 compressed domain

In this study the authors proposed a real-time video object segmentation algorithm that works in the H.264 compressed domain. The algorithm utilises the motion information from the H.264 compressed bit stream to identify background motion model and moving objects. In order to preserve spatial and temporal continuity of objects, Markov random field (MRF) is used to model the foreground field. Quantised transform coefficients of the residual frame are also used to improve segmentation result. Experimental results show that the proposed algorithm can effectively extract moving objects from different kinds of sequences. The computation time of the segmentation process is merely about 16 ms per frame for CIF size frame, allowing the algorithm to be applied in real-time applications.     

采用特征描述的H.264压缩视频运动分割

H.264视频压缩标准以其优良的压缩效率和编码灵活性得到了广泛的应用。提出了一种基于H.264压缩域的运动对象分割方法,首先从压缩视频中提取运动场,采用加权中值滤波方法滤除运动场的噪声矢量,再运用后向估计的方法重建预测运动场并进行运动场的累积,然后基于幅度、散度和旋度3个运动特征,采用改进的统计区域合并方法将运动对象分割出来。实验结果表明,该方法可有效地从H.264压缩视频中提取运动对象且分割质量较好。     

运动目标的自动空时分割算法

在视频应用中,运动目标的提取是一个重要的研究课题。为了对运动目标进行更有效的分割,提出了一种从视频序列中自动提取运动目标的空时分割算法。该算法在时域分割中采用基于齐异矢量消除的目标检测方法来获得运动目标的初始模板。通常,该初始模板具有不连续的边界和一些"孔"。为了得到较为完整的目标区域,用具有距离约束的区域生长算法来补偿初始模板。而在空域分割中,分水岭分割则通过考虑全局信息来增强其分割的精确性。然后,精确的运动目标即可通过空时融合模块提取出来。试验结果表明,该空时分割算法是有效的。     

一种压缩域中的快速运动目标提取算法

论文提出了一种工作于MPEG压缩域的快速运动目标提取算法。算法以通过部分解码得到的运动向量和亮度分量的直流DCT系数作为输入,提取P帧的运动目标。首先采用鲁棒性回归分析估计全局运动,标记出与全局运动不一致的宏块,得到运动块的分布;然后将运动向量场插值作为时间域的特征,将重构的直流图像转换到LUV颜色空间作为空间域的特征,采用快速平均移聚类找到时间和空间特征具有相似性的区域,得到细化的区域边界;最后结合运动块分布和聚类分析的结果,通过基于马尔可夫随机场的统计标号方法进行背景分离,得到运动目标的掩模。实验结果表明该算法可以有效地消除运动向量噪声的影响,并有很高的处理速度,对于CIF格式的视频码流,每秒可以处理约50帧。     

Object based watermarking for H.264/AVC video resistant to rst attacks

In this paper, a compressed domain video watermarking scheme is proposed which embeds the watermark in the homogeneous moving object within a shot of video sequence to resist geometric attacks such as rotation, scaling etc. Intuitively, object based watermarking results low payload and has the least impact on visual quality since the object area is generally small and highly textured. The proposed work has two main contributions, firstly, an existing compressed domain motion coherent block detection algorithm [7] is extended to detect the moving objects within a video shot and secondly, a watermarking scheme has been proposed by embedding within the moving objects to resist RST attacks. A comprehensive set of experiments has been carried out to justify the applicability of the proposed scheme over the existing literature.     

Mean shift clustering-based moving object segmentation in the H.264 compressed domain

This study presents a mean shift clustering-based moving object segmentation approach in the H.264 compressed domain. The motion information extracted from H.264 compressed video, including motion vectors (MVs) and partitioned block size, are used as segmentation cues. The MVs are processed by normalisation, weighted 3D median filter and motion compensation to obtain a reliable and salient MV field. The partitioned block size is used as a measure of motion texture in the process of the MV field. Based on the processed MV field, the authors employ the mean shift-based mode seeking in spatial, temporal and range domain to develop a new approach for compact representation of the MV field. Then, the MV field is segmented into different motion-homogenous regions by clustering the modes with small spatial and range distance, and each object is represented by some dominant modes. Experimental results for several H.264 compressed video sequences demonstrate good performance and efficiency of the proposed segmentation approach.     

基于改进运动历史图像的多运动目标实时跟踪

提出了一种利用视频图像对运动目标进行实时检测与跟踪的新方法.该方法利用基于改进的时间片的运动历史图像(tMHI)的灰度阶梯轮廓方法对多个运动目标进行检测,通过卡尔曼滤波器对多目标进行跟踪,并得到了各个运动目标的轨迹曲线,进而实现了对视频图像中多目标的跟踪.同时,该方法对多个目标的遮挡问题获得了明显的改善效果.实验结果表明,该方法能够对复杂场景下的多个目标进行有效的识别和准确的跟踪,系统的实时性强,识别率高,而且该方法对于复杂视频监视系统场景中的光照变化、雨雾等干扰具有较强的稳健性.     

A framework for unsupervised mesh based segmentation of moving objects

Multimedia analysis usually deals with a large amount of video data with a significant number of moving objects. Often it is necessary to reduce the amount of data and to represent the video in terms of moving objects and events. Event analysis can be built on the detection of moving objects. In order to automatically process a variety of video content in different domain, largely unsupervised moving object segmentation algorithms are needed. We propose a fully unsupervised system for moving object segmentation that does not require any restriction on the video content. Our approach to extract moving objects relies on a mesh-based combination of results from colour segmentation (Mean Shift) and motion segmentation by feature point tracking (KLT tracker). The proposed algorithm has been evaluated using precision and recall measures for comparing moving objects and their colour segmented regions with manually labelled ground truth data. Results show that the algorithm is comparable to other state-of-the-art algorithms. The extracted information is used in a search and retrieval tool. For that purpose a moving object representation in MPEG-7 is implemented. It facilitates high performance indexing and retrieval of moving objects and events in large video databases, such as the search for similar moving objects occurring in a certain period.     

Coupled object detection and tracking from static cameras and moving vehicles

We present a novel approach for multi-object tracking which considers object detection and spacetime trajectory estimation as a coupled optimization problem. Our approach is formulated in an MDL hypothesis selection framework, which allows it to recover from mismatches and temporarily lost tracks. Building upon a multi-view/multi-category object detector, it localizes cars and pedestrians in the input images. The 2D object detections are converted to 3D observations, which are accumulated in a world coordinate frame. Trajectory analysis in a spacetime window yields physically plausible trajectory candidates. Tracking is achieved by performing model selection after every frame. At each time instant, our approach searches for the globally optimal set of spacetime trajectories which provides the best explanation for the current image and all evidence collected so far, while satisfying the constraints that no two objects may occupy the same physical space, nor explain the same image pixels at any time. Successful trajectory hypotheses are then fed back to guide object detection in future frames. The resulting approach can initialize automatically and track a large and varying number of objects from both static and moving cameras. We evaluate our approach on several challenging video sequences with both a surveillance-type scenario and a scenario where the input videos are taken from a moving vehicle.     

基于熵模型的压缩域运动对象检测

随着视频编解码技术的发展,H.264已成为最主要的标准之一。为了能从H.264码流中准确有效地分割和提取出运动对象,提出了一种压缩域下的对象检测算法。该算法主要利用压缩域下对象的运动矢量信息,先对矢量进行中值滤波预处理,目的是为了减少运动估计算法和对象实际运动所产生的误差。然后利用基于熵的原理,建立运动对象在空间和时间上的一致性模型。在该模型基础上,采用最大熵方法自适应获得阈值,检测得到运动对象。实验结果证明,该算法可以获得比较好的检测结果。     

基于块划分的HEVC运动目标分割方法

针对压缩域视频的运动对象分割在复杂背景下分割精度不高的问题,提出一种基于最新压缩编码HEVC的运动分割方法。首先从HEVC压缩码流中提取块划分和相对应的运动矢量信息,并分别在帧内和帧间对运动矢量进行空域和时域的标签分类,然后利用MRF模型对标签场进行运动一致性估计,得到更精确的运动目标,最后输出MRF分割后形成的掩模信息。通过实验证明,该运动分割方法能够达到有效并可靠的分割效果,尤其对于多目标运动的视频分割效果优于其他比较的方法。     

基于EM聚类的H.264压缩域视频对象实时分割算法

从压缩域直接分割视频对象比传统的像素域分割具有快速高效的特点,目前已有不少从MPEG域分割的方法,但从H.264压缩域分割的甚少。为此提出一种基于H.264域的实时分割运动对象方法,该算法先对当前视频帧进行全局运动估计和补偿,然后对4×4的运动矢量场进行分类处理,最后对非零运动矢量使用改进的EM聚类分割算法。本文算法对多个视频序列进行了实验,结果表明,该算法针对静止背景和运动背景的视频序列都能达到较精确的实时分割。     

基于模糊聚类的H.264压缩域视频对象分割

鉴于压缩域视频运动分割方法在分割速度上的优越性,提出一种基于H.264的压缩域视频运动对象分割方法,对初始的运动矢量场进行去噪、中值滤波、校正和累积处理,得到更可靠的运动矢量场,用改进的模糊C-均值聚类算法分割出视频序列中的运动对象。实验结果表明,该方法可以快速准确地提取出视频序列中的运动对象。     

Real-time abandoned and stolen object detection based on spatio-temporal features in crowded scenes

Abandoned and stolen object detection is a challenging task due to occlusion, changes in lighting, large perspective distortion, and the similarity in appearance of different people. This paper presents real-time detection methods of abandoned and stolen objects in a complex video. The adaptive background modeling method is applied to stable tracking and the ghost image removing. To detect abandoned and stolen objects, the methods determine spatio-temporal relationship between moving people and suspicious drops. The space first detection method measures the distance between a moving object and a non-moving object in spatial change analysis. The time first detection method conducts temporal change analysis and then spatial change analysis. The potential abandoned object is classified as a definite abandoned or stolen object by two-level detection approach. The time-to-live timer is applied by adjusting several key parameters on each camera and environment. In experiments, we show the experimental results to evaluate our proposed methods using benchmark datasets.     

Robust object tracking via online dynamic spatial bias appearance models

This paper presents a robust object tracking method via a spatial bias appearance model learned dynamically in video. Motivated by the attention shifting among local regions of a human vision system during object tracking, we propose to partition an object into regions with different confidences and track the object using a dynamic spatial bias appearance model (DSBAM) estimated from region confidences. The confidence of a region is estimated to re ect the discriminative power of the region in a feature space, and the probability of occlusion. We propose a novel hierarchical Monte Carlo (HAMC) algorithm to learn region confidences dynamically in every frame. The algorithm consists of two levels of Monte Carlo processes implemented using two particle filtering procedures at each level and can efficiently extract high confidence regions through video frames by exploiting the temporal consistency of region confidences. A dynamic spatial bias map is then generated from the high confidence regions, and is employed to adapt the appearance model of the object and to guide a tracking algorithm in searching for correspondences in adjacent frames of video images. We demonstrate feasibility of the proposed method in video surveillance applications. The proposed method can be combined with many other existing tracking systems to enhance the robustness of these systems.     

Optical flow approximation based motion object extraction for MPEG-2 video stream

This paper presents a compressed-domain motion object extraction algorithm based on optical flow approximation for MPEG-2 video stream. The discrete cosine transform (DCT) coefficients of P and B frames are estimated to reconstruct DC + 2AC image using their motion vectors and the DCT coefficients in I frames, which can be directly extracted from MPEG-2 compressed domain. Initial optical flow is estimated with Black’s optical flow estimation framework, in which DC image is substituted by DC + 2AC image to provide more intensity information. A high confidence measure is exploited to generate dense and accurate motion vector field by removing noisy and false motion vectors. Global motion estimation and iterative rejection are further utilized to separate foreground and background motion vectors. Region growing with automatic seed selection is performed to extract accurate object boundary by motion consistency model. The object boundary is further refined by partially decoding the boundary blocks to improve the accuracy. Experimental results on several test sequences demonstrate that the proposed approach can achieve compressed-domain video object extraction for MPEG-2 video stream in CIF format with real-time performance.