基于视觉显著性均值漂移目标跟踪
首发时间:2015-02-06
摘要:传统的均值漂移跟踪算法,当跟踪条件复杂时无法显著区分颜色相近的目标或背景导致跟踪失败.提出基于视觉显著性均值漂移目标跟踪算法。提取目标的颜色对比度、方向和强度等多个视觉显著性特征,利用相对背景差异定义视觉显著性测度,在多个特征中选择显著性最大的特征。利用显著性测度加权融合形成多特征融合视觉显著性目标模型,并嵌入到均值漂移方法中完成目标跟踪。实验结果证明,该方法与传统跟踪算法相比具有较强的鲁棒性和较高的准确性。?????
For information in English, please click here
Meanshift robust object tracking based on visual saliency
Abstract:When the tracking condition is complex, the traditional mean shift tracking algorithm, cannot distinguish the target and background with similar color and cause tracking failure. The meanshift tracking algorithm based on visual saliency is proposed. Extracting visual saliency characteristics such as color, direction and luminance, define relative differences between object and background as visual saliency measure, select the optimal feature which is most remarkable in multiple characteristics. Multi-features fusion model is formed by significant measure weighted visual saliency features. embedded it into mean shift method to complete target tracking. The experimental results show that, this method has strong robustness and high accuracy compared with traditional tracking algorithms.
Keywords: image processing meanshift visual saliency multi-features fusion
基金:
论文图表:
引用
No.4631362100062314****
同行评议
共计0人参与
勘误表
基于视觉显著性均值漂移目标跟踪
评论
全部评论0/1000