|
|
|
| Object compressive tracking based on adaptive multi-feature appearance model |
| LU Wei1,2, XIANG Zhi-yu1,2, YU Hai-bin3, LIU Ji-lin1,2 |
| 1. Department of Information Science and Electronic Engineering, Zhejiang University, Hangzhou 310027, China; 2. Zhejiang Provincial Key Laboratory of Information Network Technology, Hangzhou 310027, China; 3. School of Electronic and Information, Hangzhou Dianzi University, Hangzhou 310018, China |
|
|
|
Abstract An adaptive multi-feature modelling method was proposed to resolve the problems of simple feature and inflexible modelling existed in the appearance model of compressive tracking. This method makes the visual representation more abundant and comprehensive through fusing intensity with the Surf-type feature which has strong power to describe the detail information like gradient and edge. A two-stage measurement matrix is constructed to measure the multi-dimension features. The Hellinger distance between a feature’s distributions of positive and negative samples is computed to analyze the feature’s ability of discriminating the object from background. The weights of features in the statistical model can be adjusted adaptively to help the model efficiently explore information that is useful for object tracking, and update according to the changes of object and background. Experimental results show that this adaptive multi-feature modelling method can describe the complex changes of object and background in the real world more accurately, and greatly improve the tracking algorithm’s robustness and precision, while holding the high efficiency.
|
|
Published: 01 December 2014
|
|
|
基于自适应多特征表观模型的目标压缩跟踪
针对压缩跟踪算法中表观模型的视觉表达特征单一、统计模型缺乏柔性的问题,提出一种自适应的多特征表观建模方法.该方法引入了对梯度、边缘等图像细节描述能力更强的Surf特征,并通过构建两级观测矩阵解决多维特征的观测问题,与亮度特征进行融合,使视觉表达更加丰富、全面；通过计算正负样本特征所服从的概率分布曲线的Hellinger距离,分析特征对目标和背景的区分能力,自适应地调整统计模型中各特征之间的权重,使统计模型能更好地利用对目标跟踪有益的信息,根据目标和背景的变化及时进行更新.实验结果表明：该自适应多特征表观模型能更加准确地描述实际场景中目标和背景的复杂变化,在保持高效率的同时,极大地提高了跟踪算法的鲁棒性和准确性.
|
|
| ［1］ LI X, HU W, SHEN C, et al. A survey of appearance models in visual object tracking ［J］. ACM transactions on Intelligent Systems and Technology (TIST), 2013, 4(4): 58-99.
［2］ DONOHO D L. Compressed sensing［J］. IEEE Transactions on Information Theory, 2006, 52(4): 1289-1306.
［3］ CANDES E J, TAO T. Near-optimal signal recovery from random projections: Universal encoding strategies？［J］. IEEE Transactions on Information Theory, 2006, 52(12): 5406-5425.
［4］ ZHANG K, ZHANG L, YANG M H. Real-time compressive tracking［C］∥Computer Vision-ECCV 2012. Berlin Heidelberg： Springer, 2012: 864-877.
［5］ 朱秋平, 颜佳, 张虎, 等. 基于压缩感知的多特征实时跟踪［J］. 光学精密工程, 2013, 21(2): 438-444.
ZHU Qiu-ping, YAN Jia, ZHANG Hu, et al. Real-time tracking using multiple features based on compressive sensing［J］. Optics and Precision Engineering, 2013, 21(2): 438-444.
［6］ RAO C R. A review of canonical coordinates and an alternative to correspondence analysis using Hellinger distance［J］. Questiió: Quaderns d’Estadística, Sistemes, Informatica i Investigació Operativa, 1995, 19(1): 23-63.
［7］ CANDES E J, TAO T. Decoding by linear programming［J］. IEEE Transactions on Information Theory, 2005, 51(12): 4203-4215.
［8］ BAY H, TUYTELAARS T, VAN GOOL L. Surf: Speeded up robust features［C］∥Computer Vision-ECCV 2006. Berlin Heidelberg: Springer, 2006: 404-417.
［9］ DALAL N, TRIGGS B. Histograms of oriented gradients for human detection［C］∥IEEE Computer Society Conference on Computer Vision and Pattern Recognition. Piscataway: IEEE, 2005, 1: 886-893.
［10］ ZHU Q, YEH M C, CHENG K T, et al. Fast human detection using a cascade of histograms of oriented gradients［C］∥ IEEE Computer Society Conference on Computer Vision and Pattern Recognition. Piscataway: IEEE, 2006, 2: 1491-1498.
［11］ COLLINS R T, LIU Y. On-line selection of discriminative tracking features［C］∥ Proceedings of Ninth IEEE International Conference on Computer Vision. Piscataway: IEEE, 2003: 346-352.
［12］ CHEN Y, PARK P S, LI A. A Novel Approach of On-line Discriminative Tracking Feature Selection［J］. International Journal of Computer and Information Technology(IJCIT), 2013, 2(3): 392-397.
［13］ 胡洁. 高维数据特征降维研究综述［J］. 计算机应用研究, 2008, 25(9): 2601-2606.
HU Jie. Survey on feature dimension reduction for high-dimensional data［J］. Application Research of Computers, 2008, 25(9): 2601-2606. |
|
Viewed |
|
|
|
Full text
|
|
|
|
|
Abstract
|
|
|
|
|
Cited |
|
|
|
|
| |
Shared |
|
|
|
|
| |
Discussed |
|
|
|
|
