高斯加权的多分类器物体追踪

doi:10.11772/j.issn.1001-9081.2014.08.2394

计算机应用 ›› 2014, Vol. 34 ›› Issue (8): 2394-2398.DOI: 10.11772/j.issn.1001-9081.2014.08.2394

高斯加权的多分类器物体追踪

兰远东¹,邓辉舫²,蔡昭权¹,杨雄¹

1. 惠州学院计算机科学系，广东惠州516007
2. 华南理工大学计算机科学与工程学院，广州510640

收稿日期:2014-02-20 修回日期:2014-04-14 出版日期:2014-08-01 发布日期:2014-08-10
通讯作者: 兰远东
作者简介:兰远东(1975-)，男，贵州桐梓人，讲师，博士，主要研究方向：模式识别、机器学习；邓辉舫(1957-)，男，湖南常德人，教授，博士，主要研究方向：智能计算、大数据处理；蔡昭权(1970-)，男，广东陆丰人，教授，硕士，主要研究方向：模式识别、图像与视频处理；杨雄(1981-)，男，湖北洪湖人，讲师，博士，主要研究方向：模式识别、智能系统。
基金资助:
国家自然科学基金资助项目;惠州市科技计划项目

Gaussian weighted multiple classifiers for object tracking

LAN Yuandong¹,DENG Huifang²,CAI Zhaoquan¹,YANG Xiong¹

1. Department of Computer Science, Huizhou University, Huizhou Guangdong 516007, China;
2. School of Computer Science and Engineering, South China University of Technology, Guangzhou Guangdong 510640, China

Received:2014-02-20 Revised:2014-04-14 Online:2014-08-01 Published:2014-08-10
Contact: LAN Yuandong
Supported by:
national natural science fund

摘要/Abstract

摘要：

针对在物体外观快速变化的情况下，大多数弱学习器不能捕获物体新的特征分布，导致追踪失败的问题，提出了高斯加权的联机多分类器增强算法。该算法为每一个领域问题定义一个弱分类器，每个弱分类器包括一个简单的视觉特征和阈值，引入高斯加权函数来权衡每个弱分类器在特定样本上的贡献，通过多分类器联合学习来提高追踪性能。在物体追踪过程中，联机多分类器在对物体定位的同时还能估计物体的姿态，能够成功地学习多模态外观模型，在物体外观快速变化的情况下追踪物体。实验结果表明：所提算法在经过一个较短序列的训练后，平均追踪错误率为12.8%，追踪性能明显提升。

Abstract:

When the appearance of an object changes rapidly, most of the weak learners can not capture the new feature distributions which will lead to tracking failure. In order to deal with that issue, a Gaussian weighted online multiple classifiers algorithm boosting for object tracking was proposed. This algorithm defined one weak classifier which included a simple visual feature and a threshold for each domain problem. Gaussian weighting function was introduced to weigh each weak classifier's contribution in a particular sample, therefore the tracking performance was improved through joint learning of multiple classifiers. In the process of object tracking, online multiple classifiers can not only simultaneously determine the location and estimate the pose of the object, but also successfully learn multi-modal appearance models and track an object under rapid appearance changes. The experimental results show that, after a short initial training phase, the average tracking error rate of the proposed algorithm is 12.8%, which proves that the tracking performance has enhanced significantly.

中图分类号:

TP391.413

兰远东邓辉舫蔡昭权杨雄. 高斯加权的多分类器物体追踪[J]. 计算机应用, 2014, 34(8): 2394-2398.

LAN Yuandong DENG Huifang CAI Zhaoquan YANG Xiong. Gaussian weighted multiple classifiers for object tracking[J]. Journal of Computer Applications, 2014, 34(8): 2394-2398.

参考文献

［1］YILMAZ A, JAVED O, SHAH M. Object tracking: a survey ［J］. ACM Journal of Computing Surveys, 2006, 38(4): 1-45.
［2］LIU Y, FAN J, WANG D W. Review of intelligent video surveillance with single camera ［C］// Proceedings of the Fourth International Conference on Machine Vision: Machine Vision, Image Processing, and Pattern Analysis. Bellingham: SPIE Press, 2012: 834923-834925.
［3］BABENKO B, YANG M H, BELONGIE S. Visual tracking with online multiple instance learning ［C］// CVPR 2009: Proceedings of the 2009 IEEE Conference on Computer Vision and Pattern Recognition. Piscataway: IEEE Press, 2009: 983-990.
［4］XIE Y, QU Y, LI C, et al. Online multiple instance gradient feature selection for robust visual tracking ［J］. Pattern Recognition Letters, 2012, 33(9): 1075-1082.
［5］SHEN D, ZHANG H, XUE Y, et al. Advances in multimedia modeling ［M］. Berlin: Springer, 2013: 517-520.
［6］ZHANG J, WANG H. A novel tracking method based on incremental semi-supervised discriminant analysis ［J］. Journal of Nanjing University: Natural Sciences, 2012, 48(4): 397-404.(张继,王洪元.一种基于增量半监督判别分析的跟踪方法［J］.南京大学学报:自然科学版,2012,48(4):397-404.)
［7］JIANG G, WANG X, SHI Z. A rule learning algorithm for event detection based on semantic trajectory ［J］. Journal of Computer Research and Development, 2012, 49(12): 2623-2630.(姜广,王晓峰,史忠植.一种基于语义轨迹的事件规则学习算法［J］.计算机研究与发展,2012,49(12):2623-2630.)
［8］LIU W, LUO Y, DAI P, et al. Visual tracking of infrared object on the sea using dense sampling features ［J］. Journal of Xiamen University: Natural Science, 2013, 52(6): 764-769.(刘伟盛,罗燕龙,戴平阳,等.基于稠密采样的海上红外目标跟踪算法［J］.厦门大学学报:自然科学版,2013,52(6):764-769.)
［9］GAO S, ZHANG H, FANG X. Multi-view canonical correlation analysis based Web spam detection ［J］. Application Research of Computers, 2013, 30(3): 810-813.(高爽,张化祥,房晓南.基于多视图典型相关分析的垃圾网页检测［J］.计算机应用研究,2013,30(3):810-813.)
［10］LADICKY L, STURGESS P, RUSSELL C, et al. Joint optimization for object class segmentation and dense stereo reconstruction ［J］. International Journal of Computer Vision, 2012, 100(2): 122-133.
［11］LIU W, DAI P, LI C. Object tracking based on online multiple instance: learning with feature weighted fusion ［J］. Computer Engineering and Applications, 2013, 48(12): 67-72.(刘薇，戴平阳，李翠华.特征加权融合的在线多示例学习跟踪算法［J］.计算机工程与应用,2013,48(12):67-72.)
［12］KIM T K, CIPOLLA R. MCBoost: multiple classifier boosting for perceptual co-clustering of images and visual features ［C］// Proceedings of the 22nd Annual Conference on Neural Information Processing Systems. Berlin: Springer, 2008: 841-856.
［13］FREUND Y, SCHAPIRE R E. A decision theoretic generalization of online learning and an application to boosting ［J］. Journal of Computer and System Sciences, 1997, 55(1): 119-139.
［14］JEBARA T S, PENTLAND A. Parameterized structure from motion for 3D adaptive feedback tracking of faces ［C］// CVPR'97: Proceedings of the 1997 Conference on Computer Vision and Pattern Recognition. Washington, DC: IEEE Computer Society, 1997: 144-150.
［15］JEREMIAH E, MARSHALL L, SISSON S A, et al. Specifying a hierarchical mixture of experts for hydrologic modeling: gating function variable selection ［J］. Water Resources Research, 2013, 49(5): 2926-2939.
［16］HAN W, ZHU J, XIANG Z, et al. An improved target tracking method based on multiple instances learning ［J］. Computer Applications and Software, 2013, 30(9): 276-279.(韩文静,朱俊平,向直扬,等.改进的基于多示例学习的目标跟踪方法研究［J］.计算机应用与软件,2013,30(9):276-279.)
［17］AUDRINO F, MEIER P. Empirical pricing kernel estimation using a functional gradient descent algorithm based on splines ［M］. St. Gallen: University of St. Gallen Press, 2012:17-29.
［18］HALL P, MARSHALL D, MARTIN R. Merging and splitting eigenspace models ［J］. IEEE Transactions on Pattern Analysis and Machine Intelligence, 2000, 22(9): 1042-1049.

[1]	温静, 宋建伟. 基于多级全局信息传递模型的视觉显著性检测[J]. 计算机应用, 2021, 41(1): 208-214.
[2]	许长青, 陈振杰, 侯仁福. 融合不精准先验知识的Landsat 8 OLI影像深度学习分类方法[J]. 计算机应用, 2020, 40(12): 3550-3557.
[3]	陈利霞, 班颖, 王学文. 基于张量核范数与3D全变分的背景减除[J]. 计算机应用, 2020, 40(9): 2737-2742.
[4]	邓茜文, 冯子亮, 邱晨鹏. 基于近红外与可见光双目视觉的活体人脸检测方法[J]. 计算机应用, 2020, 40(7): 2096-2103.
[5]	王肖, 魏嘉旺, 袁玉波. 基于特征部位圆形域的人脸图像修复方法[J]. 计算机应用, 2020, 40(3): 847-853.
[6]	刘颖, 王凤伟, 刘卫华, 艾达, 李芸, 杨凡超. 基于亮度分区模糊融合的高动态范围成像算法[J]. 计算机应用, 2020, 40(1): 233-238.
[7]	王文杰, 乔志伟, 牛蕾, 席雅睿. 自适应步长非局部全变分约束迭代图像重建算法[J]. 计算机应用, 2020, 40(1): 245-251.
[8]	严经纬, 李强, 王春茂, 谢迪, 王保青, 戴骏. 面部运动单元检测研究综述[J]. 计算机应用, 2020, 40(1): 8-15.
[9]	王书朋, 赵瑶. 基于自适应分割的多曝光图像融合算法[J]. 计算机应用, 2020, 40(1): 252-257.
[10]	马伟苹, 李文新, 孙晋川, 曹鹏霞. 基于粗精立体匹配的双目视觉目标定位方法[J]. 计算机应用, 2020, 40(1): 227-232.
[11]	王鑫, 张鑫, 宁晨. 基于多特征降维和迁移学习的红外人体目标识别方法[J]. 计算机应用, 2019, 39(12): 3490-3495.
[12]	曹昀炀, 王涛. 耦合先验拉普拉斯坐标的半监督图像分割算法[J]. 计算机应用, 2019, 39(9): 2695-2700.
[13]	王克强, 张雨帅, 王保群. 基于Retinex理论的多曝光图像融合算法[J]. 计算机应用, 2019, 39(7): 2087-2092.
[14]	李文俊, 陈斌, 李建明, 钱基德. 基于深度神经网络的表面划痕识别方法[J]. 计算机应用, 2019, 39(7): 2103-2108.
[15]	朱锴, 付忠良, 陈晓清. 基于卷积神经网络的超声图像左心室分割方法[J]. 计算机应用, 2019, 39(7): 2121-2124.

高斯加权的多分类器物体追踪

Gaussian weighted multiple classifiers for object tracking

PDF

可视化

摘要/Abstract

引用本文

使用本文

参考文献

相关文章 15

编辑推荐

Metrics