具有尺度和旋转适应性的长时间目标跟踪

Robust Long-term Object Tracking With Adaptive Scale and Rotation Estimation

目标发生尺度和旋转变化会给长时间目标跟踪带来很大的挑战,针对该问题,本文提出了具有尺度和旋转适应性的鲁棒目标跟踪算法.首先针对跟踪过程中目标存在的尺度变化和旋转运动,提出一种基于傅里叶-梅林变换和核相关滤波的目标尺度和旋转参数估计方法.该方法能够实现连续空间的目标尺度和旋转参数估计,采用核相关滤波提高了估计的鲁棒性和准确性.然后针对长时间目标跟踪过程中,有时不可避免地会出现跟踪失败的情况(例如由于长时间半遮挡或全遮挡等),提出一种基于直方图和方差加权的目标搜索方法.当目标丢失时,通过提出的搜索方法能够快速从图像中确定目标可能存在的区域,使得跟踪算法具有从失败中恢复的能力.本文还训练了两个核相关滤波器用于估计跟踪结果的置信度和目标平移,通过专门的核相关滤波器能够使得估计的跟踪结果置信度更加准确和鲁棒,置信度的估计结果可用于激活基于直方图和方差加权的目标搜索模块,并判断搜索窗口中是否包含目标.本文在目标跟踪标准数据集(Online object tracking benchmark, OTB)上对提出的算法和目前主流的目标跟踪算法进行对比实验,验证了本文提出算法的有效性和优越性.

A robust long-term object tracking algorithm with adaptive scale and rotation estimation is proposed to deal with the challenges of scale and rotation changes during long-term object tracking. Firstly, a robust scale and rotation estimation method is proposed based on the Fourier-Mellin transform and the kernelized correlation filter for scale changes and rotation of the object. Then a weighted object searching method based on histogram and variance is proposed to deal with possible tracker’s failures in the long-term tracking(due to long-term semi-occlusion or full occlusion, etc.).When the tracked object is lost, the object can be re-located in the image using the proposed searching method, so the tracker can be recovered from failures. Moreover, this paper also trains two other kernelized correlation filters to estimate the object’s translation and the confidence of tracking results. The customized kernelized correlation filter can make the estimated confidence more accurate and robust. The confidence is used to activate a weighted object searching module and determine whether the searching windows contain objects or not. The proposed algorithm is compared with the state-of-the-art tracking algorithms using the online object tracking benchmark(OTB). Experimental results validate the effectiveness and superiority of the proposed tracking algorithm.