Automatic extraction and tracking of the tongue contours

Computerized analysis of the tongue surface movement can provide valuable information to speech and swallowing research. Ultrasound technology is currently the most attractive modality for the tongue imaging mainly because of its high video frame rate. However, problems with ultrasound imaging, such as noise and echo artifacts, refractions, and unrelated reflections pose significant challenges for computer analysis of the tongue images and hence specific methods must be developed. This paper presents a system that is developed for automatic extraction and tracking of the tongue surface movements from ultrasound image sequences. The ultrasound images are supplied by the head and transducer support system (HATS), which was developed in order to fix the head and support the transducer under the chin in a known position without disturbing speech. In this work, we propose a novel scheme for the analysis of the tongue images using deformable contours. We incorporate novel mechanisms to 1) impose speech related constraints on the deformations; 2) perform spatiotemporal smoothing using a contour postprocessing stage; 3) utilize optical flow techniques to speed up the search process; and 4) propagate user supplied information to the analysis of all image frames. We tested the system's performance qualitatively and quantitatively in consultation with speech scientists. Our system produced contours that are within the range of manual measurement variations. The results of our system are extremely encouraging and the system can be used in practical speech and swallowing research in the field of otolaryngology.     

运动目标的自动分割与跟踪

该文提出了一种对视频序列中的运动目标进行自动分割的算法。该算法分析图像在L U V空间中的局部变化,同时使用运动信息来把目标从背景中分离出来。首先根据图像的局部变化,使用基于图论的方法把图像分割成不同的区域。然后,通过度量合成的全局运动与估计的局部运动之间的偏差来检测出运动的区域,运动的区域通过基于区域的仿射运动模型来跟踪到下一帧。为了提高提取的目标的时空连续性,使用Hausdorff跟踪器对目标的二值模型进行跟踪。对一些典型的MPEG-4测试序列所进行的评估显示了该算法的优良性能。     

Active contours for video object tracking using region, boundary and shape information

In this paper, we propose a robust model for tracking in video sequences with non-static backgrounds. The object boundaries are tracked on each frame of the sequence by minimizing an energy functional that combines region, boundary and shape information. The region information is formulated by minimizing the symmetric Kullback–Leibler (KL) distance between the local and global statistics of the objects versus the background. The boundary information is formulated using a color and texture edge map of the video frames. The shape information is calculated adaptively to the dynamic of the moving objects and permits tracking that is robust to background distractions and occlusions. Minimization of the energy functional is implemented using the level set method. We show the effectiveness of the approach for object tracking in color, infrared (IR), and fused color-infrared sequences.     

Nonstationary autoregressive modeling of object contours

A spatially variant circular autoregressive (SVCAR) model is introduced for the analysis and classification of closed shape boundaries. The model represents a closed shape boundary sequence as the output of a nonstationary all-pole linear system (driven by white noise) whose coefficient's spatial evolution can be expressed as a truncated function expansion. Features derived from the SVCAR model are shown to be invariant to shape scaling, rotation, and translation. A shape-matching algorithm is developed to optimally adjust the SVCAR model coefficients for changes in contour sequence starting point. Laboratory experiments involving object sets representative of industrial, military, and geographic shapes are presented. Superior classification results are demonstrated     

Tracking nonparameterized object contours in video

We propose a new method for contour tracking in video. The inverted distance transform of the edge map is used as an edge indicator function for contour detection. Using the concept of topographical distance, the watershed segmentation can be formulated as a minimization. This new viewpoint gives a way to combine the results of the watershed algorithm on different surfaces. In particular, our algorithm determines the contour as a combination of the current edge map and the contour, predicted from the tracking result in the previous frame. We also show that the problem of background clutter can be relaxed by taking the object motion into account. The compensation with object motion allows to detect and remove spurious edges in background. The experimental results confirm the expected advantages of the proposed method over the existing approaches.     

Active contours for tracking distributions

A new approach to tracking using active contours is presented. The class of objects to be tracked is assumed to be characterized by a probability distribution over some variable, such as intensity, color, or texture. The goal of the algorithm is to find the region within the current image, such that the sample distribution of the interior of the region most closely matches the model distribution. Two separate criteria for matching distributions are examined, and the curve evolution equations are derived in each case. The flows are shown to perform well in experiments.     

Automatic video object generation tool: segmentation and tracking of persons in real time

In the field of the multimedia applications, future standards as mpeg4 or mpeg7 allow new image content representation that goes far from the basic compression. This way, the objects of interest become entities that the user can manipulate. Nevertheless, the norm does not specify how objects have to be segmented, even if in some cases this is a complex task. In the framework of the European project m4m (mpeg fo(u)r Mobiles, Medea project Al 16), we have in charge the development of a demonstrator able to segment and track a speaker in videophone and vidéoconférence applications.     

基于运动跟踪匹配技术的视频对象提取

本文提出一种自动分割VOP的技术。其方法是：先对初始帧使用形态运动滤波技术提取出初始运动对象的二值轮廓模型,并在后继帧中使用豪斯道夫对象跟踪器跟踪运动以对象模型;而为了适应对象的形状变化,本文使用活动轮廓模型（snake）技术对运动心合匹配;最后根据一系列精确的二值轮廓引导提取运动对象序列。实验结果表明,我们的算法可有效地提取视频对象平面。     

Joint audio-video object localization and tracking

There has been a tremendous amount of research on object localization either involving microphone arrays or video cameras. Considerable less attention has been paid, however, to object localization and tracking based on joint audio-video processing thus far. This may be related to the lack of suitable algorithms for object localization simultaneously using multimicrophone outputs and color image sequences. In this article, we propose a solution to this problem. Before elaborating on joint audio-video processing, we review some previous work the areas of audio and video object localization. Then a recursive sensor fusion method based on decentralized Kalman filtering is introduced. Unfortunately, the decentralized Kalman filter cannot be directly used for joint audio-video object localization due to specific properties of the audio sensor. By properly adjusting the local audio position estimator, however, we manage to keep the overall architecture. We stress the general methodology     

目标实时跟踪与预测算法研究

实时成像跟踪系统要求对运动目标能够有较快的响应速度,跟踪的响应时间越短,系统的实时性就越好,从而可靠的跟踪系统显得尤为重要。文中在研究了目前常用几种跟踪算法的基础上,提出一种基于目标特征匹配和Kalman预测相结合的跟踪方法,选取目标的灰度直方图信息做为特征匹配模板,使用Kalman滤波器对目标在下一帧图像中可能出现的位置进行预测,在预测范围内进行搜索及模板匹配,实验结果表明,该跟踪算法能够对目标实现稳定可靠的跟踪。     

动态目标实时跟随及其预测算法应用研究

实时成像跟踪系统要求对运动目标能够有较快的响应速度,跟踪的响应时间越短,系统的实时性就越好,从而可靠的跟踪系统显得尤为重要。文中在研究了目前常用几种跟踪算法的基础上,提出了一种基于目标特征匹配和Kalman预测相结合的跟踪方法,选取目标的灰度直方图信息做为特征匹配模板,使用Kalman滤波器对目标在下一帧图像中可能出现的位置进行预测,在预测范围内进行搜索及模板匹配,实验结果表明,该跟踪算法能够对目标实现稳定可靠的跟踪。     

Edge-aware object pixel-level representation tracking

Recently, there has been a trend in tracking to use more refined segmentation mask instead of coarse bounding box to represent the target object. Some trackers proposed segmentation branches based on the tracking framework and maintain real-time speed. However, those trackers use a simple FCNs structure and lack of the edge information modeling. This makes performance quite unsatisfactory. In this paper, we propose an edge-aware segmentation network, which uses the complementarity between target information and edge information to provide a more refined representation of the target. Firstly, We use the high-level features of the tracking backbone network and the correlation features of the classification branch of the tracking framework to fuse, and use the target edge and target segmentation mask for simultaneous supervision to obtain an optimized high-level feature with rough edge information and target information. Secondly, we use the optimized high-level features to guide the low-level features of the tracking backbone network to generate more refined edge features. Finally, we use the refined edge features to fuse with the target features of each layer to generate the final mask. Our approach has achieved leading performance on recent pixel-wise object tracking benchmark VOT2020 and segmentation datasets DAVIS2016 and DAVIS2017 while running on 47 fps. Code is available at https://github.com/TJUMMG/EATtracker.     

具有遮挡鲁棒性的目标中心距离加权的跟踪算法

本文提出了一种实时跟踪算法,以目标中心距离加权的目标图像直方图作为模板,采用mean-shift迭代方法进行目标定位;当目标被部分遮挡时,用"分块匹配"的方法提高算法鲁棒性;对于超过一定像素的较大目标,本算法进行"降采样",大大减小运算量,从而实现了对大尺度目标的实时跟踪.实时视频流的实际跟踪系统验证了该方法的有效性.     

结合目标预估计与Mean Shift理论的运动目标跟踪算法

图像的运动包括目标、背景和平台的运动,复杂的运动关系增加了目标跟踪的难度.提出了一种有效的基于Mean Shift理论的运动目标跟踪算法.为提高算法的实时性,对Mean Shift算法的核函数进行了改进,使得加减运算替代乘方和浮点运算,大大提高了运算效率;并通过对迭代权值的改进,强化了初始模板的主要信息,提高了算法跟踪与背景相似目标的能力;采用自动更新模板的策略,克服了目标特征分布发生改变的问题;在此基础上,引入了目标预检测,提出了综合背景差分检测的运动目标跟踪算法,实验表明:该算法在目标被遮挡情况下具有较强的适应性.     

Automatic moving object and background separation

In this paper, we propose a segmentation method of reduced computational complexity aimed at separating the moving objects from the background in a generic video sequence. This task may be accomplished at the coder site to support the functionalities foreseen by new multimedia scenarios, and in particular the content-based functionalities focused by the MPEG-4 activity, allowing the user to access and decode single objects of a video sequence. The proposed algorithm discriminates between background and foreground by means of a higher-order statistics (HOS) significance test performed on a group of inter-frame differences, followed by a motion detection phase, producing a binary segmentation map. The HOS threshold is adaptively changed, based on the estimated background activity and on the potential presence of slowly moving objects. The map is refined by a final regularization stage implemented by means of a cascade of morphological filters. The algorithm performance were tested through the wide experimental activity carried out during the ISO MPEG-4 N2 Core Experiment on Automatic Segmentation Techniques, in which the authors are currently involved. Typical results obtained on MPEG4 sequences are here shown, in order to illustrate the segmentation algorithm performance.     

Active-matting-based object tracking with color cues

The linear canonical transform (LCT) has been shown to be a useful and powerful analyzing tool in optics and signal processing. Many results of this transform are already known, including its uncertainty principles (UPs). The existing UPs of the LCT for complex signals can only provide sharp bounds with LCT parameters satisfying $a_1/b_1\ne a_2/b_2$ . However, in most cases, we strive to find a lower bound, but not a sharper bound, since a lower bound often leads to optimization problems in signal processing applications. In this paper, we first present a much briefer and more transparent derivation to obtain a general uncertainty principle of the LCT for arbitrary signals via operator methods. Then, we derive lower bounds of three UPs of the LCT for complex signals, which are tighter lower bounds than the existing ones. We also prove that the derived results hold for arbitrary LCT parameters.     

Tri-tracking method for robust object tracking

A novel tri-tracking method for tracking objects is proposed. The tracker trains three support vector machines with labelled frames, in which an object is represented using three independent feature histograms. Based on the confidence map created by each classifier, the object position is given by the mean shift. Experiments on several video sequences show the robustness and accuracy of the proposed method.     

Pixel feature-weighted scale-adaptive object tracking algorithm

An effective object tracking method using weighted pixel features was proposed to deal with all kinds of complicated tracking situations,such as target movement,rotation,background interference and scaling and so on.First,the color feature and location information of the pixels in the target area were used to build the object model.Then the average weight image was used to estimate the scale variation coefficient.The aim was to adapt to the scale changes of the target.Finally,an update model was proposed,which was able to renew the object model and background model.The experimental results show that the proposed algorithm could make full use of the differences between pixels in the target area,so it can track more quickly and more effectively with strong robustness.     

Automatic detection of myocardial contours in cine-computed tomographic images

Quantitative evaluation of cardiac function from cardiac images requires the identification of the myocardial walls. This generally requires the clinician to view the image and interactively trace the contours. This method is susceptible to great variability that depends on the experience and knowledge of the particular operator tracing the contours. The particular imaging modality that is used may also add tracing difficulties. Cine-computed tomography (cine-CT) is an imaging modality capable of providing high quality cross-sectional images of the heart. CT images, however, are cluttered, i.e., objects that are not of interest, such as the chest wall, liver, stomach, are also visible in the image. To decrease this variability, investigators have developed computer-assisted or near-automatic techniques for tracing these contours. All of these techniques, however, require some operator intervention to confidently identify myocardial borders. The authors present a new algorithm that automatically finds the heart within the chest, and then proceeds to outline (detect) the myocardial contours. Information at each tomographic slice is used to estimate the contours at the next tomographic slice, thus allowing the algorithm to work in near-apical cross-sectional images where the myocardial borders are often difficult to identify. The algorithm does not require operator input and can be used in a batch mode to process large quantities of data. An evaluation and correction phase is included to allow an operator to view the results and selectively correct portions of contours. The authors tested the algorithm by automatically identifying the myocardial borders of 27 cardiac images obtained from three human subjects and quantitatively comparing these automatically determined borders with those traced by an experienced cardiologist.