摘要
为了解决目标因遮挡、跟踪框发生漂移后相关滤波跟踪算法仍持续更新目标模型和滤波器模型,导致背景信息被更新到目标模型和滤波器模型中的情况,提出一种自适应模型更新策略。采用相关滤波方法得到新的目标位置;提取新目标位置的统计协方差特征并计算其与协方差模板的相似性;根据相似性判断是否更新目标模型和滤波器模型。实验结果表明:所提方法有效解决了因目标形变、遮挡等情况导致目标模型和滤波器模型的更新问题,提高了相关滤波目标跟踪的精度。
In order to solve the problem of obscuring of target,the related filtering tracking algorithm keeps updating the target model and the filter model after the tracking frame drifts,which cause background information is updated into the target model and filter model,an adaptive model updating strategy is proposed. New target position is obtained by using the correlation filtering method. Statistical covariance feature of the new target position is extracted and its similarity with the covariance template is calculated. Judge whether update target model and filter model according to similarity or not. The experimental results show that the proposed method effectively solves the problem of target model updating and filter model updating due to target deformation and occlusion,and improves the precision of the related filtering target tracking.
引文
[1] Bolme D S,Beveridge J R,Draper B A,et al. Visual object tracking using adaptive correlation filters[C]∥Computer Vision and Pattern Recognition,IEEE,2010:2544-2550.
[2] Henriques J F,Rui C,Martins P,et al. Exploiting the circulant structure of tracking-by-detection with kernels[C]∥European Conf on Computer Vision,ECCV 2012,Berlin Heidelberg:Springer,2012:702-715.
[3] Danelljan M,Hger G,Shahbaz Khan F,et al. Accurate scale estimation for robust visual tracking[C]∥Proceedings of the British Machine Vision Conference,2014:65. 1-65. 11.
[4] Danelljan M,Khan F S,Felsberg M,et al. Adaptive color attributes for real-time visual tracking[C]∥IEEE Conference on Computer Vision and Pattern Recognition,IEEE,2014:1090-1097.
[5] Li Y,Zhu J. A scale adaptive kernel correlation filter tracker with feature integration[C]∥2014,8926:254-265.
[6] Galoogahi H K,Sim T,Luccey S. Correlation filters with limited boundaries[C]∥Proceedings of the IEEE Conference on Computer Vision and Pattern Recognition,2014:4630-4638.
[7] Henriques J F,Caseiro R,Martins P,et al. High-speed tracking with kernelized correlation filters[J]. IEEE Transactions on Pattern Analysis and Machine Intelligence 2015,37(3):583-596.
[8]赵翱东,石磊,刘飞.基于时空特征的实时相关滤波目标跟踪[J].传感器与微系统. 2016,35(8):119-122,126.
[9] Asha C S,Narasimhadhan A V. Adaptive learning rate for visual tracking using correlation filters[J]. Procedia Computer Science,2016,89:614-622.
[10] Wang M,Liu Y,Huang Z. Large Margin Object Tracking with circulant feature maps[J]. ar Xiv preprint ar Xiv. 2017,1703.05020.
[11] Tuzel O,Porikli F,Meer P. Region covariance:A fast descriptor for detection and classification[C]∥. Computer Vision ECCV2006,2006:589-600.
[12] Wu Y,Lim J,Yang M H. Online object Tracking:A benchmark[C]∥Computer Vision and Pattern Recognition,IEEE,2013:2411-2418.
[13] Zhang K,Zhang L,Yang M H. Real-time compressive tracking[C]∥European Conference on Computer Vision. Springer-Verlag,2012:864-877.
[14] Hare S,Golodetz S,Saffari A,et al. Struck:Structured output tracking with kernels[J]. IEEE Transactions on Pattern Analysis and Machine Intelligence,2016,38(10):2096-2109.
[15] Kalal Z,Mikolajczyk K,Matas J. Tracking-learning-detection[J].IEEE Transactions on Pattern Analysis and Machine Intelligence,2012,34(7):1409-1422.