摘要
为了解决单目标跟踪中的光照变化、部分遮挡问题,提出了一种结构化的加权联合特征表观模型.该模型将被跟踪的目标图像划分成若干图像块,在每个图像块内计算其颜色特征和纹理特征,将这些特征加权形成特征向量作为目标的表观模型.以该模型为基础,利用贝叶斯理论,提出一种跟踪方法.实验结果表明了该方法的有效性.
A structural appearance model with weighted associated features is proposed to deal with illumination variation and partial occlusion questions in single object tracking. The tracked object image is divided into small image blocks. Thereafter, the color features and textural features are calculated within each block. Next, these features are weighted and a vector is composed, which is presumed the appearance model of the tracked object. Subsequently, through the application of Bayes＇ theorem, a tracking method based on the appearance model is proposed. Finally, the effectiveness of the proposed tracking method is demonstrated through experimental results.
出处
《信息与控制》
CSCD
北大核心
2015年第3期372-378,384,共8页
Information and Control
基金
国家自然科学基金面上项目(NSFC 61375014)
