A template consensus method for visual tracking

Visual tracking is a challenging problem in computer vision. Recently, correlation filter-based trackers have shown to provide excellent tracking performance. Inspired by a sample consensus approach proposed for foreground detection, which classifies a given pixel as foreground or background based on its similarity to recently observed samples, we present a template consensus tracker based on the kernelized correlation filter (KCF). Instead of keeping only one target appearance model in the KCF, we make a feature pool to keep several target appearance models in our method and predict the new target position by searching for the location of the maximal value of the response maps. Both quantitative and qualitative evaluations are performed on the CVPR2013 tracking benchmark dataset. The results show that our proposed method improves the original KCF tracker by 8.17% in the success plot and 8.11% in the precision plot.     

A multiprocessor-oriented visual tracking system

The design and prototypal realization of a visual tracking system is presented. The approach to target identification is nonconventional, in that it relies on an architecture composed of multiple standard neural networks (multilayer perceptrons) and exploits the information contained in simple features extracted from images, performing a small number of operations. Therefore, the tracking functions are learned by examples, rather than implemented directly. The system demonstrates that a quite complex task such as visual target tracking can be easily obtained by a suitable neural architecture. The fast tracking algorithm and the parallel structure allow a true real-time operation. The system exploits a two-level neural-network hierarchy with a number of parallel networks and an “arbiter”. The training set consists of various geometrical shapes, preprocessed to yield the data vectors. The experimental hardware implementation is based on multiple processing units, implementing the neural architecture, and serves as a prototype for the analysis of the system in practice. A small-sized realization can also be obtained     

Fragmentation handling for visual tracking

Object detection and tracking using background subtraction suffers from the fragmentation problem which means one object fragments into several blobs because of being similar with the reference image in color. In this paper, we build a visual tracking framework using background subtraction for object detection, and we address the association difficulty of blobs with objects caused by the fragmentation problem by two steps. We firstly cluster the blobs according to the boundary distances of them estimated by an approximating method proposed in this paper. Blobs clustered into the same blob-set are considered from the same object. Secondly, we consider blob-sets possibly from the same object if they exhibit coherent motion, since blobs of the same object may be clustered into different blob-sets if the object fragments severely. A background-matching method is proposed to determine whether two blob-sets exhibiting coherent motion are truly from the same object or from different objects. We test the proposed methods on several real-world video sequences. Quantitative and qualitative experimental results show that the proposed methods handle the problems caused by fragmentation effectively.     

Eye tracking — A new interface for visual exploration

Eye-tracking technology offers a natural and immediate way of communicating human intentions to a computer. Eye movements reflect interests and may be analysed to drive computer functionality in games, image and video search, and other visual tasks. This paper examines current eye tracking technologies and their applications. Experiments are described that show that target images can be identified more rapidly by eye tracking than using a mouse interface. Further results show that an eye-tracking technology provides an efficient interface for locating images in a large database. Finally the paper speculates about how the technology may enter the mass market as costs decrease.     

A self-organizing approach to background subtraction for visual surveillance applications.

Detection of moving objects in video streams is the first relevant step of information extraction in many computer vision applications. Aside from the intrinsic usefulness of being able to segment video streams into moving and background components, detecting moving objects provides a focus of attention for recognition, classification, and activity analysis, making these later steps more efficient. We propose an approach based on self organization through artificial neural networks, widely applied in human image processing systems and more generally in cognitive science. The proposed approach can handle scenes containing moving backgrounds, gradual illumination variations and camouflage, has no bootstrapping limitations, can include into the background model shadows cast by moving objects, and achieves robust detection for different types of videos taken with stationary cameras. We compare our method with other modeling techniques and report experimental results, both in terms of detection accuracy and in terms of processing speed, for color video sequences that represent typical situations critical for video surveillance systems.     

一种视觉跟踪中的模板更新策略

针对复杂场景中的目标外观和背景变化引起的模板 更新问题, 提出了一种新的视觉跟踪模板更新策略,用以提高目标模板正确性。算法利用特征信息在 时间和空间上的区别和变化,进行特征子量分类更新,避免了模型过更新,提高了目标模型 的容错能力,使更新带来的误差尽量小,以适应目标和背景信息的不断变化,在一定程度上 提高了跟踪算法的精准度和鲁棒性。实验结果表明,本文方法在视频跟踪系统中具有优越的 性 能,可以在目标运动、变化和遮挡情况下实现鲁棒跟踪。     

A visual analysis/synthesis feedback loop for accurate face tracking

We propose a novel approach for face tracking, resulting in a visual feedback loop: instead of trying to adapt a more or less realistic artificial face model to an individual, we construct from precise range data a specific texture and wireframe face model, whose realism allows the analysis and synthesis modules to visually cooperate in the image plane, by directly using 2D patterns synthesized by the face model. Unlike other feedback loops found in the literature, we do not explicitly handle the 3D complex geometric data of the face model, to make real-time manipulations possible. Our main contribution is a complete face tracking and pose estimation framework, with few assumptions about the face rigid motion (allowing large rotations out of the image plane), and without marks or makeup on the user's face. Our framework feeds the feature-tracking procedure with synthesized facial patterns, controlled by an extended Kalman filter. Within this framework, we present original and efficient geometric and photometric modelling techniques, and a reformulation of a block-matching algorithm to make it match synthesized patterns with real images, and avoid background areas during the matching. We also offer some numerical evaluations, assessing the validity of our algorithms, and new developments in the context of facial animation. Our face-tracking algorithm may be used to recover the 3D position and orientation of a real face and generate a MPEG-4 animation stream to reproduce the rigid motion of the face with a synthetic face model. It may also serve as a pre-processing step for further facial expression analysis algorithms, since it locates the position of the facial features in the image plane, and gives precise 3D information to take into account the possible coupling between pose and expressions of the analysed facial images.     

A biomimetic VLSI sensor for visual tracking of small moving targets

Taking inspiration from the visual system of the fly, we describe and characterize a monolithic analog very large-scale integration sensor, which produces control signals appropriate for the guidance of an autonomous robot to visually track a small moving target. This sensor is specifically designed to allow such tracking even from a moving imaging platform which experiences complex background optical flow patterns. Based on relative visual motion of the target and background, the computational model implemented by this sensor emphasizes any small-field motion which is inconsistent with the wide-field background motion.     

Computational sensor for visual tracking with attention

This paper presents a VLSI embodiment of an optical tracking computational sensor which focuses attention on a salient target in its field of view. Using both low-latency massive parallel processing and top-down sensory adaptation, the sensor suppresses interference front features irrelevant for the task at hand, and tracks a target of interest at speeds of up to 7000 pixels/s. The sensor locks onto the target to continuously provide control for the execution of a perceptually guided activity. The sensor prototype, a 24×24 array of cells, is built in 2-μm CMOS technology. Each cell occupies 62 μm×62 μm of silicon, and contains a photodetector and processing electronics     

Rotation-aware correlation filters for robust visual tracking

Recent years have witnessed several modified discriminative correlation filter (DCF) models exhibiting excellent performance in visual tracking. A fundamental drawback to these methods is that rotation of the target is not well addressed which leads to model deterioration. In this paper, we propose a novel rotation-aware correlation filter to address the issue. Specifically, samples used for training of the modified DCF model are rectified when rotation occurs, rotation angle is effectively calculated using phase correlation after transforming the search patch from Cartesian coordinates to the Log-polar coordinates, and an adaptive selection mechanism is further adopted to choose between a rectified target patch and a rectangular patch. Moreover, we extend the proposed approach for robust tracking by introducing a simple yet effective Kalman filter prediction strategy. Extensive experiments on five standard benchmarks show that the proposed method achieves superior performance against state-of-the-art methods while running in real-time on single CPU.     

Transferring visual prior for online object tracking

Visual prior from generic real-world images can be learned and transferred for representing objects in a scene. Motivated by this, we propose an algorithm that transfers visual prior learned offline for online object tracking. From a collection of real-world images, we learn an overcomplete dictionary to represent visual prior. The prior knowledge of objects is generic, and the training image set does not necessarily contain any observation of the target object. During the tracking process, the learned visual prior is transferred to construct an object representation by sparse coding and multiscale max pooling. With this representation, a linear classifier is learned online to distinguish the target from the background and to account for the target and background appearance variations over time. Tracking is then carried out within a Bayesian inference framework, in which the learned classifier is used to construct the observation model and a particle filter is used to estimate the tracking result sequentially. Experiments on a variety of challenging sequences with comparisons to several state-of-the-art methods demonstrate that more robust object tracking can be achieved by transferring visual prior.     

End-to-end DeepNCC framework for robust visual tracking

In this paper, we propose an NCC-based object tracking deep framework, which can be well initialized with the limited target samples in the first frame. The proposed framework contains a pretrained model, online feature fine-tuning layers and tracking processes. The pretrained model provides rich feature representations while online feature fine-tuning layers select discriminative and generic features for the tracked object. We choose normalized cross-correlation as a template tracking layer to perform the tracking process. To enable the learned features representation closely coordinated to the tracked target, we jointly train the feature representation network and tracking processes. In online tracking, an adaptive template and a fixed template are fused to find the optimal tracking results. Scale estimation and a high-confidence model update scheme are perfectly integrated into the framework to adapt to the target appearance changes. The extensive experiments demonstrate that the proposed tracker achieves superior performance compared with other state-of-the-art trackers.     

Data fusion for visual tracking with particles

The effectiveness of probabilistic tracking of objects in image sequences has been revolutionized by the development of particle filtering. Whereas Kalman filters are restricted to Gaussian distributions, particle filters can propagate more general distributions, albeit only approximately. This is of particular benefit in visual tracking because of the inherent ambiguity of the visual world that stems from its richness and complexity. One important advantage of the particle filtering framework is that it allows the information from different measurement sources to be fused in a principled manner. Although this fact has been acknowledged before, it has not been fully exploited within a visual tracking context. Here we introduce generic importance sampling mechanisms for data fusion and discuss them for fusing color with either stereo sound, for teleconferencing, or with motion, for surveillance with a still camera. We show how each of the three cues can be modeled by an appropriate data likelihood function, and how the intermittent cues (sound or motion) are best handled by generating proposal distributions from their likelihood functions. Finally, the effective fusion of the cues by particle filtering is demonstrated on real teleconference and surveillance data.     

Correlation,Kalman filter and adaptive fast mean shift based heuristic approach for robust visual tracking

Correlation tracker is computation intensive (if the search space or the template is large), has template drift problem, and may fail in case of fast maneuvering target, rapid changes in its appearance, occlusion suffered by it and clutter in the scene. Kalman filter can predict the target coordinates in the next frame, if the measurement vector is supplied to it by a correlation tracker. Thus, a relatively small search space can be determined where the probability of finding the target in the next frame is high. This way, the tracker can become fast and reject the clutter, which is outside the search space in the scene. However, if the tracker provides wrong measurement vector due to the clutter or the occlusion inside the search region, the efficacy of the filter is significantly deteriorated. Mean-shift tracker is fast and has shown good tracking results in the literature, but it may fail when the histograms of the target and the candidate region in the scene are similar (even when their appearance is different). In order to make the overall visual tracking framework robust to the mentioned problems, we propose to combine the three approaches heuristically, so that they may support each other for better tracking results. Furthermore, we present novel method for (1) appearance model updating which adapts the template according to rate of appearance change of target, (2) adaptive threshold for similarity measure which uses the variable threshold for each forthcoming image frame based on current frame peak similarity value, and (3) adaptive kernel size for fast mean-shift algorithm based on varying size of the target. Comparison with nine state-of-the-art tracking algorithms on eleven publically available standard dataset shows that the proposed algorithm outperforms the other algorithms in most of the cases.     

A NEW MODELING AND FILTERING APPROACH FOR TRACKING MANEUVERING TARGETS

A new modeling and filtering approach for tracking maneuvering targets is presented in thispaper.The approach,which makes optimal estimate for the model With the random variable possible,depends on random step modeling of target maneuvers.In the new model,the unknown targetacceleration is treated as a random variable and then estimated directly.A detector is designed tofind out the target maneuvers and the estimation algorithm will be restarted when the maneuvers oc-cur.Combination of three-dimention Kalman filter with a detector forms a tracker for maneuveringtargets.The new tracking scheme is easy to implement and its capability is illustrated in two trackingexamples in which the new approach is compared with Mooses'on the performance.     

Affine hull based target representation for visual tracking

Handling appearance variations is a challenging issue in visual tracking. Existing appearance models are usually built upon a linear combination of templates. With such kind of representation, accurate visual tracking is not desirable when heavy appearance variations are in presence. Under the framework of particle filtering, we propose a novel target representation for tracking. Namely, the target candidates are represented by affine combinations of a template set, which leads to better capability in describing unseen target appearances. Additionally, in order to adapt this representation to dynamic contexts across a video sequence, a novel template update scheme is presented. Different from conventional approaches, the scheme considers both the importance of one template to a target candidate in the current frame and the recentness of the template that is kept in the template set. Comprehensive experiments show that the proposed algorithm achieves superior performances in comparison with state-of-the-art works.     

Online semi-supervised compressive coding for robust visual tracking

In this paper we propose an online semi-supervised compressive coding algorithm, termed SCC, for robust visual tracking. The first contribution of this work is a novel adaptive compressive sensing based appearance model, which adopts the weighted random projection to exploit both local and discriminative information of the object. The second contribution is a semi-supervised coding technique for online sample labeling, which iteratively updates the distributions of positive and negative samples during tracking. Under such a circumstance, the pseudo-labels of unlabeled samples from the current frame are predicted according to the local smoothness regularizer and the similarity between the prior and the current model. To effectively track the object, a discriminative classifier is online updated by using the unlabeled samples with pseudo-labels in the weighted compressed domain. Experimental results demonstrate that our proposed algorithm outperforms the state-of-the-art tracking methods on challenging video sequences.     

Heuristic algorithm for visual tracking of deformable objects

Many vision problems require fast and accurate tracking of objects in dynamic scenes. These problems can be formulated as exploration problems and thus can be expressed as a search into a state space based approach. However, these problems are hard to solve because they involve search through a space of transformations corresponding to all the possible motion and deformation. In this paper, we propose a heuristic algorithm through the space of transformations for computing target 2D motion. Three features are combined in order to compute efficient motion: (1) a quality of function match based on a holistic similarity measurement, (2) Kullback–Leibler measure as heuristic to guide the search process and (3) incorporation of target dynamics into the search process for computing the most promising search alternatives. Once 2D motion has been calculated, the result value of the quality of function match computed is used with the purpose of verifying template updates. A template will be updated only when the target object has evolved to a transformed shape dissimilar with respect to the actual shape. Also, a short-term memory subsystem is included with the purpose of recovering previous views of the target object. The paper includes experimental evaluations with video streams that illustrate the efficiency and suitability for real-time vision based tasks in unrestricted environments.     

Robust ambiguous target handling for visual object tracking

This paper presents a robust method of handling ambiguous targets (partial occlusion, split region or mixed state of the partial occlusion and the split region) for visual object tracking. The object model is a combination of bounding box features and expected object region. These object properties are very compact and allow us to track objects in a cluttered environment. The target state is classified in the first stage by a state classifier. The state classifier is defined from a weighted cross-correlation of normalized area and normalized distance which are defined from the comparison of background model- and the motion-based object detections. The correlation can categorize the target state by using the overlap quantity of the detected objects from the both object detections. If the target is merged state (partial occlusion), we will identify and track each object in the merged region by the bounding box features. If the targets are the split region, these regions are identified and grouped by the expected object region. If the target is the mixed state, we use the methods for handling the split and the merged region. Finally, experimental results show that the proposed method can deal with tracking in cluttered environments.     

Feature compression: A framework for multi-view multi-person tracking in visual sensor networks

Visual sensor networks (VSNs) consist of image sensors, embedded processors and wireless transceivers which are powered by batteries. Since the energy and bandwidth resources are limited, setting up a tracking system in VSNs is a challenging problem. In this paper, we present a framework for human tracking in VSNs. The traditional approach of sending compressed images to a central node has certain disadvantages such as decreasing the performance of further processing (i.e., tracking) because of low quality images. Instead, we propose a feature compression-based decentralized tracking framework that is better matched with the further inference goal of tracking. In our method, each camera performs feature extraction and obtains likelihood functions. By transforming to an appropriate domain and taking only the significant coefficients, these likelihood functions are compressed and this new representation is sent to the fusion node. As a result, this allows us to reduce the communication in the network without significantly affecting the tracking performance. An appropriate domain is selected by performing a comparison between well-known transforms. We have applied our method for indoor people tracking and demonstrated the superiority of our system over the traditional approach and a decentralized approach that uses Kalman filter.