期刊
SOFT COMPUTING
卷 27, 期 14, 页码 9787-9800出版社
SPRINGER
DOI: 10.1007/s00500-023-07820-x
关键词
Visual tracking; Object extraction; Template extraction; Template matching; Feature extraction; Feature update; Motion estimation
In this paper, a visual object tracking framework based on object appearance feature update is proposed. The tracking model utilizes template spatial information and object features to track the target object in successive frames, and adapts to appearance variations by updating the tracked template feature vector. Experimental results demonstrate its high precision (84.5%) and tracking speed (10.1 fps), outperforming conventional trackers in qualitative analysis.
In computer vision applications, visual object tracking is a complex task in which object appearances change in the presence of illumination variation, occlusion, in-plane rotation, and fast motion. In the state-of-the-art approaches, trackers deal with the common model to address appearance variations with coexisting challenges. However, this approach is ineffective when dealing with simultaneous challenges because the object's features differ due to the appearance variations. To alleviate these limitations, in this paper, a visual object tracking framework that relies on an object appearance feature update is proposed. The appearance tracking model was developed using templated spatial information and object features. The tracked object template is identified by comparing the tracked template from the previous frame with the directional templates in the current frame. To adapt to appearance variations, the proposed tracking model updates the tracked template feature vector, the motion parameters, and the spatial information as it tracks the target object in successive frames. The experimental results on challenging video sequences in object tracking benchmarks demonstrate that the proposed tracking model can track objects with a precision of 84.5% at a 10.1 fps tracking speed. The qualitative analysis shows that the proposed tracking model outperforms the related conventional trackers.
作者
我是这篇论文的作者
点击您的名字以认领此论文并将其添加到您的个人资料中。
推荐
暂无数据