Randomized Ensemble Tracking

Bai, Qinxun; Wu, Zheng; Sclaroff, Stan; Betke, Margrit; Monnier, Camille

doi:10.1109/iccv.2013.255

Cited by 91 publications

(38 citation statements)

References 19 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…It is a key task to derive a proper loss function in our multi-expert framework. One straightforward option, which is in the same spirit of many ensemble based tracking methods [16,7,4], is to base the loss function (or weighting function) on the likelihood of the experts, in other words, how well the experts fit the labeled training samples. However, for online model-free tracking, training samples are labeled by the tracker.…”

Section: Expert Selection For Tracking Using Entropy Minimizationmentioning

confidence: 99%

“…Some approaches are designed to detect tracking failures and occlusions, to avert bad updates [4,22,32]. Others employ machine learning methods that are robust to sample labeling errors [3,8].…”

Section: Related Workmentioning

confidence: 99%

“…In [2,7], weak classifiers are combined and updated by a boosting algorithm for model-free tracking. The formulation of [4] combines weak SVM classifiers via randomized weighting vectors. In [3], multiple instance learning is used to avoid the error-prone, hard-labeling process.…”

Section: Related Workmentioning

confidence: 99%

See 2 more Smart Citations

MEEM: Robust Tracking via Multiple Experts Using Entropy Minimization

Zhang

Sclaroff

2014

Lecture Notes in Computer Science

Self Cite

932

767

View full text Add to dashboard Cite

Abstract. We propose a multi-expert restoration scheme to address the model drift problem in online tracking. In the proposed scheme, a tracker and its historical snapshots constitute an expert ensemble, where the best expert is selected to restore the current tracker when needed based on a minimum entropy criterion, so as to correct undesirable model updates. The base tracker in our formulation exploits an online SVM on a budget algorithm and an explicit feature mapping method for efficient model update and inference. In experiments, our tracking method achieves substantially better overall performance than 32 trackers on a benchmark dataset of 50 video sequences under various evaluation settings. In addition, in experiments with a newly collected dataset of challenging sequences, we show that the proposed multi-expert restoration scheme significantly improves the robustness of our base tracker, especially in scenarios with frequent occlusions and repetitive appearance variations.

show abstract

Section: Expert Selection For Tracking Using Entropy Minimizationmentioning

confidence: 99%

Section: Related Workmentioning

confidence: 99%

See 1 more Smart Citation

MEEM: Robust Tracking via Multiple Experts Using Entropy Minimization

Zhang

Sclaroff

2014

Lecture Notes in Computer Science

Self Cite

932

767

View full text Add to dashboard Cite

show abstract

“…With better performance, the feature integration model is becoming more welcomed than the single feature based model in recent years. The typical work include the weighted sum model [13], hierarchical model [14], the HMM model [15], the Gaussian mixture model [16] and the pyramid model [17]. The idea of integration has also been realized in the Deep Learning framework: a STCT method [18] extracted an effective feature map via convolutional networks, the networks are trained from a large scale image classification dataset for tracking.…”

Section: Introductionmentioning

confidence: 99%

Decontaminate feature for tracking: adaptive tracking via evolutionary feature subset

Liu

Wang²,

Yin³

et al. 2017

J. Electron. Imaging

View full text Add to dashboard Cite

This version is available at https://strathprints.strath.ac.uk/62675/ Strathprints is designed to allow users to access the research output of the University of Strathclyde. Unless otherwise explicitly stated on the manuscript, Copyright © and Moral Rights for the papers on this site are retained by the individual authors and/or other copyright owners. Please check the manuscript for details of any other licences that may have been applied. You may not engage in further distribution of the material for any profitmaking activities or any commercial gain. You may freely distribute both the url (https://strathprints.strath.ac.uk/) and the content of this paper for research or private study, educational, or not-for-profit purposes without prior permission or charge.Any correspondence concerning this service should be sent to the Abstract. Although various visual tracking algorithms have been proposed in the last 2-3 decades, it remains a challenging problem for effective tracking with fast motion, deformation, occlusion et al. Under complex tracking conditions, most tracking models are not discriminative and adaptive enough. When the combined feature vectors are inputted to the visual models, this may lead to redundancy caused low efficiency and ambiguity caused poor performance. In this paper, an effective tracking algorithm is proposed to decontaminate features for each video sequence adaptively, where the visual modeling is treated as an optimization problem from the perspective of evolution. Every feature vector is compared to a biological individual and then decontaminated via classical evolutionary algorithms.With the optimized subsets of features, "Curse of Dimensionality" has been avoided whilst the accuracy of the visual model has been improved. The proposed algorithm has been tested on several publicly available datasets with various tracking challenges and benchmarked with a number of state-of-the-art approaches. The comprehensive experiments have demonstrated the efficacy of the proposed methodology.

show abstract

“…Compressive sensing [24] and sparse representation [16] methods are also explored to build appearance models for object tracking. Moreover, to avoid wrong updates, multi-models [5,8,14,20,23] which have more diversities were employed for visual tracking. Large scale experiments with various evaluation criteria to gauge the state-of-the-art are given in [21].…”

Section: Introductionmentioning

confidence: 99%

Learn++ for Robust Object Tracking

Zheng

Shao

Racić

2014

Proceedings of the British Machine Vision Conference 2014

View full text Add to dashboard Cite

In this paper, a Learn++ (LPP) tracker is proposed to efficiently select specific classifiers for robust and long-term object tracking. In contrast to previous online methods, LPP tracker dynamically maintains a set of basic classifiers which are trained sequentially without accessing original data but preserving the previously acquired knowledge. The different subsets of basic classifiers can be specified to solve different sub-problems occurred in a non-stationary environment. Thus, an optimal classifier can be approximated in an active subspace spanned by selected adaptive basic classifiers. As a result, LPP tracker can address the "concept drift", by automatically adjusting the active subset and searching the optimal classifier in an active subspace spanned by the subset according to the distribution of the samples and recent performance. Experimental results show that LPP tracker yields state-of-the-art performance under various challenging environmental conditions and, especially, can overcome several challenges simultaneously.

show abstract

Randomized Ensemble Tracking

Abstract: We

Cited by 91 publications

References 19 publications

MEEM: Robust Tracking via Multiple Experts Using Entropy Minimization

MEEM: Robust Tracking via Multiple Experts Using Entropy Minimization

Decontaminate feature for tracking: adaptive tracking via evolutionary feature subset

Learn++ for Robust Object Tracking

Contact Info

Product

Resources

About