结构先验约束下的实时多目标跟踪-计算机工程与应用.pdfVIP

174 2016 ，52（4 ） Computer Engineering and Applications 计算机工程与应用 结构先验约束下的实时多目标跟踪 1 1 2 1 1 赵其鹏 ，孙永宣 ，洪 艳 ，姚婷婷 ，谢 昭 1 1 2 1 1 ZHAO Qipeng , SUN Yongxuan , HONG Yan , YAO Tingting , XIE Zhao 1.合肥工业大学 计算机与信息学院，合肥 230009 2.合肥工业大学 信息与网络中心，合肥 230009 1.School of Computer and Information, Hefei University of Technology, Hefei 230009, China 2.Center of Information and Network, Hefei University of Technology, Hefei 230009, China ZHAO Qipeng, SUN Yongxuan, HONG Yan, et al. Real-time multi-object tracking with structure priors. Computer Engineering and Applications, 2016, 52 （4 ）：174-180. Abstract ：In general, the multi-target, multi-frame data association needs to be exploited in multiple objects tracking. However, large number of samples and time complexity of the objective function limits its application. A novel real-time multi-object tracking algorithm is proposed. An on-line updating structure priors model is presented to capture the data space association. For the inference of space confidence, a dense sampling method is presented to overcome the imprecision of target state caused by traditional sparse sampling strategy. The surrounding regions of target are exacted as positive ex- ample, and then transformed circularly as negatives. Using the well-established theory of Circulant matrices, a fast learning method is achieved to infer the confidence for all possible position of target with the Fast Fourier Transform. Experimental results on a number of challenging sequences demonstrate the precision and efficiency of the method. Key words ：object tracking; structure priors; dense sampling; fast fourier transform 摘 要：多目标跟踪算法通常需要计算多帧、多目标间的数据关联，由于目标样本数量大，优化过程十分耗时，因此 往往实际应用受限。提出一种实时的多目标跟踪算法，通过建立在线更新的结构先验模型约束目标间的空间位置 关系，从而捕获多帧多目标间的数据相关性；在推理目标的空间置信度时，为克服传统方法使用稀疏采样造成样本 不足引起目标状态估计不准确的问题，采用一种新的思路：提取目标及其周围区域作为正例样本，在