Visual tracking via robust multitask sparse prototypes

Zhang, Huanlong; Hu, Shiqiang; Yu, Junyang

doi:10.1117/1.jei.24.2.023025

Cited by 18 publications

(25 citation statements)

References 0 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…Frag [27], IVT [28], MIL [29], OAB [30], L1 [31], VTS [32], TLD [33], MTT [34], CT [35], VTD [36], Struck [37] and PartT [38]. All these trackers apply different representation or inference models.…”

Section: Resultsmentioning

confidence: 99%

Multi-Dimension Support Vector Machine Based Crowd Detection and Localisation Framework for Varying Video Sequences

Mahalakshmi¹,

Kanthavel²

2016

View full text Add to dashboard Cite

In this paper, we propose a novel method for anomalous crowd behaviour detection and localization with divergent centers in intelligent video sequence through multiple SVM (support vector machines) based appearance model. In multi-dimension SVM crowd detection, many features are available to track the object robustly with three main features which include 1) identification of an object by gray scale value, 2) histogram of oriented gradients (HOG) and 3) local binary pattern (LBP). We propose two more powerful features namely gray level co-occurrence matrix (GLCM) and Gaber feature for more accurate and authenticate tracking result. To combine and process the corresponding SVMs obtained from each features, a new collaborative strategy is developed on the basis of the confidence distribution of the video samples which are weighted by entropy method. We have adopted subspace evolution strategy for reconstructing the image of the object by constructing an update model. Also, we determine reconstruction error from the samples and again automatically build an update model for the target which is tracked in the video sequences. Considering the movement of the targeted object, occlusion problem is considered and overcome by constructing a collaborative model from that of appearance model and update model. Also if update model is of discriminative model type, binary classification problem is taken into account and overcome by collaborative model. We run the multi-view SVM tracking method in real time with subspace evolution strategy to track and detect the moving objects in the crowded scene accurately. As shown in the result part, our method also overcomes the occlusion problem that occurs frequently while objects under rotation and illumination change due to different environmental conditions.

show abstract

Section: Resultsmentioning

confidence: 99%

Multi-Dimension Support Vector Machine Based Crowd Detection and Localisation Framework for Varying Video Sequences

Mahalakshmi¹,

Kanthavel²

2016

View full text Add to dashboard Cite

show abstract

“…In Mei et al [2], an efficient l 1 tracker with minimum error bound and occlusion detection is proposed. Zhang et al [14] investigate convex mixed norm l p,q (i.e. p ≥ 1, q ≥ 1) to enforce joint sparsity for the particles.…”

Section: Sparse Representation For Object Trackingmentioning

confidence: 99%

“…To explore these hidden structures, [14] employed multi task sparse learning to impose joint sparsity between the particles (tasks) yields a more accurate representation for the ensemble of particles where the sparse representation matrix X = [x 1 , . .…”

Section: Joint Sparse Modelmentioning

confidence: 99%

Enhanced Laplacian Group Sparse Learning with Lifespan Outlier Rejection for Visual Tracking

Bozorgtabar

Goecke

2015

Computer Vision -- ACCV 2014

View full text Add to dashboard Cite

Abstract. Recently, sparse based learning methods have attracted much attention in robust visual tracking due to their effectiveness and promising tracking results. By representing the target object sparsely, utilising only a few adaptive dictionary templates, in this paper, we introduce a new particle filter based tracking method, in which we aim to capture the underlying structure among the particle samples using the proposed similarity graph in a Laplacian group sparse framework, such that the tracking results can be improved. Furthermore, in our tracker, particles contribute with different probabilities in the tracking result with respect to their relative positions in a given frame in regard to the current target object location. In addition, since the new target object can be well modelled by the most recent tracking results, we prefer to utilise the particle samples that are highly associated to the preceding tracking results. We demonstrate that the proposed formulation can be efficiently solved using the Accelerated Proximal method with just a small number of iterations. The proposed approach has been extensively evaluated on 12 challenging video sequences. Experimental results compared to the state-of-the-art methods demonstrate the merits of the proposed tracker.

show abstract

“…Generative methods use appearance models to represent the target object and search for the most similar image regions to the generative model. Popular generative trackers include eigentracker [7], incremental tracker [26,23], sparse trackers [24,34,5], visual tracking decomposition [22], and so on. A drawback of these methods is that they are not designed to distinguish between target and background patches, and are prone to drift.…”

Section: Related Workmentioning

confidence: 99%

“…To design a robust tracking algorithm in spite of partial occlusion, researchers have developed sophisticated appearance models through statistical analysis [15], robust statistics [1,10], model analysis [12], learning occlusion with likelihoods [21], and sparse representation [24,35,34,5]. Among them, methods part-wisely modeling object appearance [1,13,9,27,18,29,25] become more popular partially because of their favorable property of robustness against partial occlusion.…”

Section: Introductionmentioning

confidence: 99%

Partial Occlusion Handling for Visual Tracking via Robust Part Matching

Zhang

Jia

et al. 2014

2014 IEEE Conference on Computer Vision and Pattern Recognition

112

View full text Add to dashboard Cite

Part-based visual tracking is advantageous due to its robustness against partial occlusion. However, how to effectively exploit the confidence scores of individual parts to construct a robust tracker is still a challenging problem. In this paper, we address this problem by simultaneously matching parts in each of multiple frames, which is realized by a locality-constrained low-rank sparse learning method that establishes multi-frame part correspondences through optimization of partial permutation matrices. The proposed part matching tracker (PMT) has a number of attractive properties. (1) It exploits the spatial-temporal localityconstrained property for robust part matching. (2) It matches local parts from multiple frames jointly by considering their low-rank and sparse structure information, which can effectively handle part appearance variations due to occlusion or noise. (3) The proposed PMT model has the inbuilt mechanism of leveraging multi-mode target templates, so that the dilemma of template updating when encountering occlusion in tracking can be better handled. This contrasts with existing methods that only do part matching between a pair of frames. We evaluate PMT and compare with 10 popular state-of-the-art methods on challenging benchmarks. Experimental results show that PMT consistently outperform these existing trackers.

show abstract

Visual tracking via robust multitask sparse prototypes

Cited by 18 publications

References 0 publications

Multi-Dimension Support Vector Machine Based Crowd Detection and Localisation Framework for Varying Video Sequences

Multi-Dimension Support Vector Machine Based Crowd Detection and Localisation Framework for Varying Video Sequences

Enhanced Laplacian Group Sparse Learning with Lifespan Outlier Rejection for Visual Tracking

Partial Occlusion Handling for Visual Tracking via Robust Part Matching

Contact Info

Product

Resources

About