摘要
提出一种基于在线综合直方图特征评估的运动目标跟踪算法.首先,通过融合颜色直方图和梯度方向直方图,形成一种新的综合直方图特征集,在物体的表达过程中有效融合和强化物体的颜色和轮廓描述;其次,为了实现长时间的稳定跟踪以及适应跟踪过程中物体和背景的连续变化,提出了一种新的跟踪物体权值评估算法,使可信特征在跟踪中起到更大的作用.复杂背景下的实验验证了该算法的有效性.
An online combined feature evaluation method for visual object tracking was proposed. FirstLy, a feature set was built by combining color histogram （HC） bins with gradient orientation histogram （HOG） bins, emphasizing the color and contour representation. Then a feature confidence evaluation approach was proposed to make features of larger confidences play more important roles in tracking, which ensured that the tracking could adapt to the appearance changes of either foreground or background. Experiment results verify the effectiveness and efficiency of the proposed method when objects go across complicated backgrounds.
基金
国家自然科学基金(60672147)资助
关键词
特征评估
综合直方图
目标跟踪
feature evaluation
combined feature histogram
object tracking
