基于时空分割融合的视频对象提取改进算法

提出一种基于时空分割融合的视频对象提取改进算法：时间分割基于变化检测，其关键是通过直方图分析选取合适的阈值；空间分割采用修正的分水岭变换，并通过一种基于时间距离和空间距离测度的区域合并算法克服“过分割”问题。实验结果表明，它可取得比COST211 AM更好的分割结果。     

基于时空混沌的图像分割算法

为了实现图像的自适应多级分割,利用混沌动力学系统对初值的敏感性和时空混沌中全局耦合映像混沌同步的特点,提出了一种新的基于时空混沌的图像分割算法.新算法对每个灰度等级构建全局耦合映像,建立灰度等级间关系矩阵和灰度等级更新模型,通过全局耦合映像混沌同步过程完成灰度等级的自适应聚类.实验结果表明:算法对中间类灰度等级聚类具有很好的鲁棒性,分割过程不需要人工干预,并且具有较强的灵活性,优于阈值分割和边缘检测的图像分割方法.     

基于时空信息的自动视频对象分割算法

提出一种在通用视频序列中从背景里分离出运动对象的方法.首先,使用全局运动估计和补偿、场景变化检测技术进行预处理.然后,使用四阶统计显著性检测方法从帧差图像中分离出前景和背景,并使用连通成分标记算法和一连串形态学开启、闭合操作得到修正后的二值模板图.接着,使用对称差分技术消除覆盖\显露的背景以及部分噪声.最后,使用模板匹配和更新,不仅能够得到快速变化的对象,而且能够得到视频对象暂时停止运动的部分.     

Motion segmentation by multistage affine classification

We present a multistage affine motion segmentation method that combines the benefits of the dominant motion and block-based affine modeling approaches. In particular, we propose two key modifications to a recent motion segmentation algorithm developed by Wang and Adelson (1994). 1) The adaptive k-means clustering step is replaced by a merging step, whereby the affine parameters of a block which has the smallest representation error, rather than the respective cluster center, is used to represent each layer; and 2) we implement it in multiple stages, where pixels belonging to a single motion model are labeled at each stage. Performance improvement due to the proposed modifications is demonstrated on real video frames.     

Speech segmentation and spoken document processing

Progress in both speech and language processing has spurred efforts to support applications that rely on spoken rather than written language input. A key challenge in moving from text-based documents to such spoken documents is that spoken language lacks explicit punctuation and formatting, which can be crucial for good performance. This article describes different levels of speech segmentation, approaches to automatically recovering segment boundary locations, and experimental results demonstrating impact on several language processing tasks. The results also show a need for optimizing segmentation for the end task rather than independently.     

Motion parameter recovery and 3D scene segmentation

Motion parameter recovery for AGVs based on landmarks (or corners) has been widely used in stereo navigation. Existing methods use weighted least-squares algorithms to recover motion parameters, but these estimates usually yield a zigzagged vehicle trajectory. Based on the idea that vehicle motion must be constrained by the vehicle dynamics, a new method of recovering motion parameters is introduced, resulting in considerable improvement in positional accuracy. The method can also be used for scene segmentation     

Proposals from binary tree and spatio-temporal tunnel for temporal segmentation of rough videos

Existing temporal segmentation methods suffer from the problems of high computational complexity and complicated steps. To address this issue, we present a method that combines the binary tree and spatio-temporal tunnel (STT) for temporal segmentation of rough videos. First, we compute initial cumulative spatio-temporal flow to determine flow overflow of sub-video which is divided from a rough video. Second, the decision tree is generated by combining binary tree and balance factor to dynamically adjust the sampling line of the STT. Finally, pixels on the sampling line are extracted to generate an adaptive STT for temporal proposals. Experimental results show that the computational complexity of the proposed method is significantly better than that of the comparison methods while ensuring accuracy.     

用基于运动的区域生长法分割序列图像

针对基于对象的视频编码应用,提出了一种基于运动的区域生长分割方案,将图像分割成具有一致运动特征的区域。方案中采用块匹配法进行运动估计,利用重构式开－闭及区域生长型水岭算法等形态工具进行空间域分割,以改进时间域分割。此外,为了求取可靠的运动矢量,还提出了一种变化检测算法及预选多个候选运动矢量的运动估计方法。初步的模拟结果表明,该方案是可行的。其主要优点是算法简单、计算量少。     

结合梯度重建和非线性处理的语义图像分割

语义图像分割是进一步图像分析与理解的重要基础,本文面向语义分割,针对传统分水岭分割算法对噪声敏感、易于产生过分割等问题,研究了一种结合梯度重建和非线性处理的图像分割方案.首先提出了基于形态学的开闭二次重建获得形态梯度图像的方法,然后对其进行阈值变换,以抑制噪声引起的明暗细节,之后进行给定尺度等级的非线性划分,以降低目标和背景内部细密纹理的影响,而后,采用了一种改进的分水岭浸没标记算法进行分割.实验结果表明,与传统算法和Matlab工具箱中算法相比,本文算法获得了更为理想的分割结果.     

Communication architecture for processing spatio-temporal continuous queries in sensor networks

Wireless sensor networks have revolutionized distributed micro-sensing because of their ease of deployment, ad hoc connectivity and cost-effectiveness. They have also enabled collecting and monitoring data from a very large area or possibly several independent areas geographically separated from each other and such a process is known as spatio-temporal data monitoring. In this paper, we define an energy-aware routing infrastructure that enables distributed query processing and supports processing of spatio-temporal queries within the network. As operator execution demands high computation capability, we propose a possible use of a heterogeneous sensor network where query operators are assigned to sparsely-deployed resource-rich nodes within a dense network of low power sensor nodes. We have designed an adaptive, decentralized, low communication overhead algorithm to determine optimal operator placement on the resource-rich nodes such that data transfer cost in the network is minimized. To the best of our knowledge, this is the first attempt to build an energy-aware communication architecture to enable in-network processing of spatio-temporal queries.     

Automatic segmentation of video object planes in MPEG-4 based on spatio-temporal information

Segmentation of semantic Video Object Planes (VOP‘s) from video sequence is a key to the standard MPEG-4 with content-based video coding. In this paper, the approach of automatic Segmentation of VOP‘s Based on Spatio-Temporal Information (SBSTI) is proposed. The proceeding results demonstrate the good performance of the algorithm.     

基于形态学处理与模式识别的手部图像分割

为了更有效地分割手部图像,提出一种基于彩色阈值变换、形态学处理与模式识别的手部识别算法。利用手部与周围背景在RGB颜色分量上存在一定的差异,先通过RGB阈值变幻将彩色图像二值化,然后进行闭运算与孔洞填充,最后根据周长面积比去除背景中的孤立点集以提取手部。实验证明,该算法能够较为有效地将手部从背景中提取出来。     

The processing element design for a large-scale spatio-temporal pattern clustering system

The clustering of spatio-temporal patterns are essential for many applications. Established from the biological analogy of the cortex, the parametrically coupled logistic map network (PCLMN) provides a viable solution to the clustering problem. To engineer for a single-chip spatio-temporal pattern clustering system, the highly modular PCLMN is designed in analog circuit. In this paper, the 0.6 μm 5 V CMOS design of the processing element is presented. The analog design employs self-calibration techniques to improve the accuracy and robustness of the nonlinear circuits. A fabricated element covers a die area of 0.55 mm², and consumes 240 mW power at 5 V supply. After calibration, simulation and testing results show that the element fulfills the system-level requirement of the Cort-X model for driving signals up to 1 MHz.     

Model selection in spatio-temporal electromagnetic source analysis

Several methods [model selection procedures (MSPs)] to determine the number of sources in electroencephalogram (EEG) and magnetoencphalogram (MEG) data have previously been investigated in an instantaneous analysis. In this paper, these MSPs are extended to a spatio-temporal analysis if possible. It is seen that the residual variance (RV) tends to overestimate the number of sources. The Akaike information criterion (AIC) and the Wald test on amplitudes (WA) and the Wald test on locations (WL) have the highest probabilities of selecting the correct number of sources. The WA has the advantage that it offers the opportunity to test which source is active at which time sample.     

复杂场景下基于动态纹理的运动分割和背景估计方法

针对现有的运动分割和背景估计方法无法分割停止运动的对象、不适用于复杂动态场景等不足,首先提出一种基于动态纹理(DT)的背景-前景混合模型(FBM),实现动态场景下前景和背景的联合表示。FBM包括一组关于位置的DT成份和一组全局DT成份,前者用于模拟本地背景运动,后者用于模拟持续性的前景运动。其次,提出一种可学习FBM参数的EM算法及变分近似策略,使得FBM在不需要人工选择阈值和不需要单独训练视频的前提下,实现多种运动复杂场景下的前景运动分割,并检测出停止运动的对象。仿真实验结果表明,与当前最新的运动分割和背景估计方法相比,该方法可显著提升背景估计和运动分割的精度。     

Group-based spatio-temporal video analysis and abstraction using wavelet parameters

In this paper, we present a spatio-temporal event-based approach to video signal analysis and abstraction employing wavelet transform features. The video signal is assumed to be a sequence of overlapping independent visual components called events, which typically are temporally overlapping compact functions that describe temporal evolution of a given set of the spatial parameters of the video signal. We utilize event-based temporal decomposition technique to resolve the overlapping arrangement of the video signal that is known to be one of the main concerns in video analysis via conventional frame-based schemes. In our method, a set of spatial parameters, extracted from the video, is expressed as a linear combination of a set of temporally overlapping compact functions, called events, through an optimization process. First, to reduce computational complexity, the video sequence is divided into overlapped groups. Next, Generalized Gaussian Density (GGD) parameters, extracted from 2D wavelet transform subbands, are used as the spatial parameters. Temporal decomposition is then applied to the GGD parameters, structured as a frame-based matrix of GGD vectors, to compute the event functions and associated orthogonal GGD parameters. Frames located at event centroids, which are much smaller in number than the number of frames in the original video, are taken as candidates for the keyframes that are selected based on a distance criterion in the feature space. Our contribution is that this still image video abstraction scheme does not need shot or cluster boundary detection, unlike current methods. Experimental results confirm the efficiency and accuracy of our approach.     

Unsupervised segmentation of low clouds from infrared METEOSATimages based on a contextual spatio-temporal labeling approach

The early and accurate segmentation of low clouds during the night-time is an important task for nowcasting. It requires that observations can be acquired at a sufficient time rate as provided by the geostationary METEOSAT satellite over Europe. However, the information supplied by the single infrared METEOSAT channel available by night is not sufficient to discriminate between low clouds and ground during night from a single image. To tackle this issue, the authors consider several sources of information extracted from an infrared image sequence. Indeed, they exploit both relevant local motion-based measurements, intensity images and thermal parameters estimated over blocks, along with local contextual information. A statistical contextual labeling process in two classes, involving "low clouds" and "clear sky," is performed on the warmer pixels. It is formulated within a Bayesian estimation framework associated with Markov random field (MRF) models. This comes to minimize a global energy function comprising three terms: two data-driven terms (thermal and motion-based ones) and a regularization term expressing a priori knowledge on the label field (expected spatial contextual properties). The authors propose a progressive minimization procedure of this energy function starting from initial reliably labeled pixels and involving only local computation     

数字图像处理中的图像分割技术及其应用

图像分割是数字图像处理中最为困难的任务之一,到目前为止仍没有适用于所有图像的通用图像分割方法.从图像分割技术的基本原理出发,比较全面地总结了有代表性的图像分割方法,在此基础之上,给出了在机车检测领域、生物医学领域和遥感图像领域中这些图像分割方法的应用,最后对图像分割的发展趋势给予了展望.     

基于Y形阵的空时二维波达方向估计

提出了一种适用于Y形阵的空时二维波达方向估计方法,在低信噪比下实现了无模糊的二维波达方向估计。该方法利用了Y形阵的特性,具有较好的稳定性和较小的估计方差。仿真结果证明了本方法的有效性。     

Image segmentation and approximation through surface type labellingand region merging

The authors present a two-stage image segmentation and approximation method for an image coding application. The proposed method is based on the surface fitting of an intensity image by a second-order polynomial function. Experimental results show its much improved performance compared with the method of Kocher-Leonardi (1986)