Tracking Algorithm with Correlation Filtering Based on Combination of Adaptive Feature Representation and Fern Classifier

doi:10.7652/xjtuxb201906014

您当前的位置：

首页 >

文章列表页 >

Tracking Algorithm with Correlation Filtering Based on Combination of Adaptive Feature Representation and Fern Classifier

更新时间：2025-07-09

- Tracking Algorithm with Correlation Filtering Based on Combination of Adaptive Feature Representation and Fern Classifier
- Vol. 53, Issue 6, Pages: 101-108(2019)
- 作者机构：
  
  1. 西安交通大学电子与信息工程学院,西安,710049
  2. 中国科学院西安光学精密机械研究所,西安,710119
  3. 中国科学院大学,北京,100049
- 作者简介：
- 基金信息：
- DOI：10.7652/xjtuxb201906014
  CLC： TP391.41
- Online First：10 June 2019，
  
  Published：2019
- 稿件说明：
移动端阅览
Tracking Algorithm with Correlation Filtering Based on Combination of Adaptive Feature Representation and Fern Classifier[J]. 2019, 53(6): 101-108.
DOI：

Tracking Algorithm with Correlation Filtering Based on Combination of Adaptive Feature Representation and Fern Classifier[J]. 2019, 53(6): 101-108. DOI： 10.7652/xjtuxb201906014.

摘要

为解决基于核相关滤波架构的跟踪算法所采用的线性插值模型更新策略无法应对目标外观突变的问题

提出一种结合自适应特征选择和蕨类分类器的目标跟踪算法(DRDCF)。首先对提取的多层目标特征层进行主成分分析降维以抽取有用的特征层; 其次

采用每帧均对模板固定更新的进取型滤波器结合满足门限条件才进行更新的保守型滤波器定位目标

将进取型滤波器用于预测目标的下一帧位置

将保守型滤波器用于计算进取型滤波器以及检测器产生的预测位置的可靠性; 当进取型滤波器预测位置不可靠时

检测器产生预测位置

最后通过对比两者预测位置的可靠性择优确定目标最佳预测位置。实验结果表明

DRDCF算法可以有效解决目标突变所造成的模型污染以及跟踪失败问题

在OTB2015数据集上精度及覆盖率两项指标相较于结合通道和空间约束的相关滤波算法分别提升了2.78%和4.26%

达到前沿算法的效果。

Abstract

A target tracking algorithm based on combination of adaptive feature representation and a fern classifier is proposed to solve the problem that the linear interpolation update strategy used in the tracking algorithms based on kernel correlation filter architecture cannot deal with sudden changes in target's appearance. Firstly

dimension reduction through principal component analysis is carried out for extracted feature layers of the target to obtain good feature layers. Secondly

the target location is determined by combination of the passive-aggressive filters

where the aggressive filter is updated with a fixed update rate in each frame and the passive filter is updated only when a threshold condition is met. The aggressive filter is used to predict the next frame position of the target

and the passive filter is used to calculate the reliability of the predicted position generated by the aggressive filter and detector. When the prediction of the aggressive filter is unreliable

the detector predicts a position of the target. The best prediction of the target position is finally determined by comparing the reliability of both predicted positions. Experiment results show that the proposed method effectively addresses the model corruption and tracking failure caused by target mutation. Results on the OTB2015 dataset show that the proposed method significantly outperforms the discriminative correlation filter with channel and spatial reliability algorithm by 2.78% and 4.26% in precision and area under curve metrics

respectively.

关键词

Keywords

references

SUN Xin, CHEUNG N M, YAO Hongxun, et al. Non-rigid object tracking via deformable patches using shape-preserved KCF and level sets [C]∥16th IEEE International Conference on Computer Vision. Piscataway, NJ, USA: IEEE, 2017: 5496-5504.

SUN Xin, YAO Hongxun, ZHANG Shengping, et al. Non-rigid object contour tracking via a novel supervised level set model [J]. IEEE Transactions on Image Processing, 2015, 24(11): 3386-3399.

初红霞, 谢忠玉, 王科俊. 一种结合颜色纹理直方图的改进型Camshift目标跟踪算法 [J]. 西安交通大学学报, 2018, 52(3): 145-152.

CHU Hongxia, XIE Zhongyu, WANG Kejun. An improved camshift target tracking algorithm based on joint color texture histogram [J]. Journal of Xi'an Jiaotong University, 2018, 52(3): 145-152.

杨源, 库涛, 查宇飞, 等. 快速多特征金字塔的尺度目标跟踪方法 [J]. 西安交通大学学报, 2016, 50(10): 49-56.

YANG Yuan, KU Tao, ZHA Yufei, et al. Fast multi-feature pyramids for scale-adaptive object tracking [J]. Journal of Xi'an Jiaotong University, 2016, 50(10): 49-56.

韩玉兰, 朱洪艳, 韩崇昭. 采用随机矩阵的多扩展目标滤波器 [J]. 西安交通大学学报, 2015, 49(7): 98-104.

HAN Yulan, ZHU Hongyan, HAN Chongzhao. A multi-target filter based on random matrix [J]. Journal of Xi'an Jiaotong University, 2015, 49(7): 98-104.

DONG Xingping, SHEN Jianbing. Occlusion-aware real-time object tracking [J]. IEEE Transactions on Multimedia, 2017, 19(4): 763-771.

GALOOGAHI H K, FAGG A, LUCEY S. Learning background-aware correlation filters for visual tracking [C]∥16th IEEE International Conference on Computer Vision. Piscataway, NJ, USA: IEEE, 2017: 1144-1152.

MUELLER M, SMITH N, GHANEM B. Context-aware correlation filter tracking [C]∥30th IEEE Conference on Computer Vision and Pattern Recognition. Piscataway, NJ, USA: IEEE, 2017: 1387-1395.

DANELLJAN M, HAGER G, KHAN F S, et al. Adaptive decontamination of the training set: a unified formulation for discriminative visual tracking [C]∥29th IEEE Conference on Computer Vision and Pattern Recognition. Piscataway, NJ, USA: IEEE, 2016: 1430-1438.

WANG Ning, ZHOU Wengang, LI Houqiang, Reliable re-detection for long-term tracking [J]. IEEE Transactions on Circuits and Systems for Video Technology, 2019, 29(3): 730-743.

BOLME D S, BEVERIDGE J R, DRAPER B A, et al. Visual object tracking using adaptive correlation filters [C]∥2010 IEEE Computer Society Conference on Computer Vision and Pattern Recognition. Piscataway, NJ, USA: IEEE, 2010: 2544-2550.

DALAL N, TRIGGS B. Histograms of oriented gradients for human detection [C]∥2005 IEEE Computer Society Conference on Computer Vision and Pattern Recognition. Piscataway, NJ, USA: IEEE, 2005: 886-893.

DANELLJAN M, KHAN F S, FELSBERG M, et al. Adaptive color attributes for real-time visual tracking [C]∥27th IEEE Conference on Computer Vision and Pattern Recognition. Piscataway, NJ, USA: IEEE, 2014: 1090-1097.

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

LIANG Pengpeng, BLASCH E, LING Haibin. Encoding color information for visual tracking: algorithms and benchmark [J]. IEEE Transactions on Image Processing, 2015, 24(12): 5630-5644.

ZABIH R, WOODFILL J. Non-parametric local transforms for computing visual correspondence [M]∥Lecture Notes in Computer Science: Vol 801. Berlin, Germany: Springer-Verlag, 1994: 151-158.

DANELLJAN M, HAGER G, KHAN F S, et al. Accurate scale estimation for robust visual tracking [C]∥25th British Machine Vision Conference. Durham, UK: BMVA, 2014: 20145200381236.

LUKEZIC A, VOJIR T, CEHOVIN L. Discriminative correlation filter with channel and spatial reliability [C]∥30th IEEE Conference on Computer Vision and Pattern Recognition. Piscataway, NJ, USA: IEEE, 2017: 4847-4856.

BERTINETTO L, VALMADRE J, GOLODETZ S, et al. Staple: complementary learners for real-time tracking [C]∥29th IEEE Conference on Computer Vision and Pattern Recognition. Piscataway, NJ, USA: IEEE, 2016: 1401-1409.

MA Chao, YANG Xiaokang, ZHANG Chongyang, et al. Long-term correlation tracking [C]∥2015 IEEE Computer Society Conference on Computer Vision and Pattern Recognition. Piscataway, NJ, USA: IEEE, 2015: 5388-5396.

Views

下载量

CSCD

Alert me when the article has been cited

提交

Tools

Publicity Resources

Tracking and Locating Algorithm of 3-Dimensional Moving Targets Using 2-Dimensional Angular Measurements from Multiple Passive Sensors

A Fixed-Interval Smoothing Algorithm for Linear Systems with General Correlated Noises

New Parameter Estimation Method for Flying Target Based on Bearing Measurements

A New Parameter Estimation Method for Flying Target Using Bearing Measurement

Data Association Method Based on Regression Analysis

Related Author

刘哲 1

卫军胡 1

赵军 3

姚东升 3

孙国基 1

Related Institution

Mailbox 21 Ext. 187

State Key Laboratory for Manufacturing System Engineering, Xi'an Jiao tong University

Systems Engineering Institute, Xi'an Jiaotong University

China Satellite Maritime Tracking and Control Department, Jiangyin

Urumqi Mailbox 21 Ext. 187

⁰