结合PLS表示与随机梯度的目标优化跟踪

金广智; 石林锁; 刘浩; 牟伟杰; 蔡艳平

doi:10.11999/JEIT151082

结合PLS表示与随机梯度的目标优化跟踪

doi: 10.11999/JEIT151082

基金项目:

国家自然科学基金(61501470)

计量
- 文章访问数: 1182
- HTML全文浏览量: 136
- PDF下载量: 328
- 被引次数: 0
出版历程
- 收稿日期: 2015-09-23
- 修回日期: 2016-05-10
- 刊出日期: 2016-08-19

Object Optimization Tracking via PLS Representation and Stochastic Gradient

Funds:

The National Natural Science Foundation of China (61501470)

摘要

摘要: 针对实际视觉跟踪中目标表观与前背景的非线性变化，论文提出一种基于偏最小二乘分析(PLS)表示与随机梯度的目标优化跟踪方法。该方法将目标跟踪转化为表示误差与分类损失的联合优化问题。首先，为了提高算法对前背景表观变化的稳定性，利用PLS理论的非线性对目标区域的前背景信息进行表达，并通过空间聚类构造多个线性外观模型来描述目标区域的动态变化，建立带约束条件的表观特征库；然后，提出一种确定性搜索机制，构造联合优化目标函数，使表示误差与分类损失最小化；结合表观建模特点，构建随机梯度分类器，对模型进行增量特征更新，最终实现对目标的稳定准确跟踪。经多场景对比实验验证，该算法能有效应对目标前背景的多种复杂变化。
- 目标跟踪 /
- 偏最小二乘 /
- 表观模型 /
- 随机梯度 /
- 联合优化
Abstract: In order to improve the stability and accuracy of the object tracking under nonlinear conditions, an object tracking algorithm based on Partial Least Squares (PLS) representation and stochastic gradient object optimization tracking is proposed. In this method, object tracking is defined as an optimization task that minimizes the representation error and classification loss. Firstly, it expresses object appearance and background information by PLS theory, learns multiple low dimensional and discriminative subspaces to describe the nonlinear appearance changes of the object. Then, a joint optimization objective function based on deterministic search mechanism is proposed. Furthermore, an stochastic gradient classifier based on incremental features updating is proposed, and make sure that it can achieve a stable tracking. Experiments show favorable performance of the proposed algorithm on sequences where the targets undergo a variety complex changes on foreground and background.
- Object tracking /
- Partial Least Squares (PLS) /
- Appearance model /
- Stochastic gradient /
- Joint optimization

HTML全文

参考文献(16)

YANG H, SHAO L, ZHENG F, et al. Recent advances and trends in visual tracking: A review[J]. Neurocomputing, 2011, 74(18): 3823-3831.

ROSS D, LIM J, LIN R S, et al. Incremental learning for robust visual tracking[J]. International Journal of Computer Vision, 2008, 77(3): 125-141.

ZHANG L and VANDER MATTEN L J P. Preserving structure in model-free tracking[J]. IEEE Transactions on Pattern Analysis and Machine Intelligence, 2014, 36(4): 756-769.

ORON S, BAR-HILLEL A, LEVI D, et al. Locally orderless tracking[C]. 2012 IEEE Conference on Computer Vision and Pattern Recognition, Rhode Island, 2012: 1940-1947.

ZHANG S, YAO H, ZHOU H, et al. Robust visual tracking based on online learning sparse representation[J]. Neurocomputing, 2013, 100(1): 31-40.

POLING B, LEMAN G, and SZLAM A. Better feature tracking through linear subspace constraints[C]. 2014 IEEE Conference on Computer Vision and Pattern Recognition (CVPR), Columbus, Ohio, 2014: 3454-3461.

GRABNER H, LEISTNER C, and BISCHOF H. Semi- supervised on-line boosting for robust tracking[C]. European Conference on Computer Vision, Crete Greece, 2010: 234-247.

BABENKO B, YANG M H, and BELONGIE S. Visual tracking with online multiple instance learning[C]. IEEE Conference on Computer Vision and Pattern Recognition, Colorado, 2011: 983-990.

KALAL Z, MATAS J, and MIKOLAJCZYK K. P-N learning: Bootstrapping binary classifiers by structural constraints[C]. IEEE Conference on Computer Vision and Pattern Recognition, California, 2010: 49-56.

WANG Qing, CHEN Feng, XU Wenli, et al. Online discriminative object tracking with local sparse representation[J]. IEEE Workshop on the Applications of Computer Vision, 2012, 12(4): 425-432.

CHEN Feng, WANG Qing, WANG Song, et al. Object tracking via appearance modeling and sparse representation [J]. Image and Vision Computing, 2013, 29(11): 787-796.

ROSIPAL R and KRAMER N. Overview and recent advances in partial least squares[J]. Latent Structure and Feature Selection, 2010, 18(3): 34-51.

HU W, LI W, ZHANG X, et al. Single and multiple object tracking using a multi-feature joint sparse representation[J]. IEEE Transactions on Pattern Analysis and Machine Intelligence, 2015, 37(4): 816-833.

ZOU H and HASTIE T. Regularization and variable selection via the elastic net[J]. Journal of the Royal Statistical Society: Series B, 2011, 67(2): 301-320.

BORDERS A, BOTTOU L, and GALLINARI P. Sgd-qn: careful quasi-newton stochastic gradient descent[J]. The Journal of Machine Learning Research, 2014, 98(10): 1737-1754.

WU Y, LIM J, and YANG M H. Object tracking benchmark [J]. IEEE Transactions on Pattern Analysis and Machine Intelligence, 2015, 37(9): 1834-1848.

施引文献

资源附件(0)

访问统计