LUO Shaoye,LIU Lisang.Multi-face tracking algorithm based on feature combination[J].Journal of Yanbian University,2018,44(01):63-68.
基于特征组合的多人脸跟踪算法
- Title:
- Multi-face tracking algorithm based on feature combination
- Keywords:
- multi-face tracking; feature combination; Camshift; ORB; Kalman filter
- 分类号:
- TP391.41
- 文献标志码:
- A
- 摘要:
- 提出了一种融合肤色特征、ORB特征和运动状态估计的多人脸跟踪算法.该算法以多线程跟踪为基础,根据不同跟踪算法的适用特点,在未受肤色干扰时依靠基于分块加权的改进Camshift算法跟踪,在受干扰时则结合包含尺度变化的ORB特征匹配算法进行跟踪.算法同时利用Kalman滤波器修正跟踪误差,以提高跟踪效果.实验表明,基于特征组合的多人脸跟踪算法具有较好的跟踪准确性和实时性.
- Abstract:
- In this paper, a multi-face tracking algorithm which integrates skin-color feature, ORB feature and motion state estimation is proposed. This algorithm creates an independent tracking thread which is based on multi-threading. And according to the characteristics of application of different tracking algorithms, it tracked the faces relied on the improved Camshift algorithm tracking which was based on block weighting with no interference of skin color. However, when disturbed by the skin color, this algorithm tracked the faces by using ORB feature matching algorithm with scale change. And it uses the Kalman filter to improve the effect of tracking and correct the tracking error. The experiment shows that the multi-face tracking algorithm based on feature combination has better tracking accuracy and higher real-time performance.
参考文献/References:
[1] Bradski G R. Computer vision face tracking as a component of a perceptual user interface[C]//Proceedings of IEEE Workshop Applications of Computer Vision. Princeton, NT: IEEE, 1998:214-219.
[2] Kass M, Witkin A, Terzopoulous D. Snake: active contour moduels[J]. International Journal of Computer Vision, 1987,1(4):321-331.
[3] Lowe D G. Object recognition from local scale-invariant features[C]//Proceeding of the 7th IEEE International Conference on Computer Vision. Corfu, Greece: IEEE, 1999.
[4] Bay H, Tuytelaars T, Van Gool L. Surf: Speeded up Robust Features[M]. Berlin Heidelberg: Springer, 2006.
[5] Rublee E, Rabaud V, Konolige K, et al. ORB: an efficient alternative to SIFT or SURF[C]//IEEE International Conference on Computer Vision. Barcelona, Spain: IEEE, 2011.
[6] Kalman R E. A new approach to linear filtering and prediction problems[J]. Transaction of the ASME-Journal of Basic Engineering, 1960,82(Series D):35-45.
[7] Michael Isard, Andrew Blake. Condensation-conditional density propagation for visual tacking[J]. International Journal of Computer Vision, 1998,29(1):5-28.
[8] Duffner S, Odobez J M. A track creation and deletion framework for long-term online multi-face tracking[J]. IEEE Transactions on Image Processing, 2013,22(1):272-285.
[9] Choi W P, Lam K M. An effective shape-texture weighted algorithm for multi-view face tracking in videos[J]. Congress on Image & Signal Processing, 2008,4:156-160.
[10] 稂龙亚,钱雪忠.基于SURF特征点的多人脸跟踪方法研究[J].计算机应用与软件,2015,32(2):178-181.
[11] 孟繁静.基于视频的实时多人脸检测跟踪与优选方法研究[D].沈阳:东北师范大学,2016.
[12] 杨超,蔡晓东,王丽娟,等.一种改进的Camshift跟踪算法及人脸检测框架[J].计算机工程与科学,2016,38(9):1863-1869.
[13] 骆绍烨,刘丽桑.基于改进Camshift的人脸跟踪算法[J].延边大学学报(自然科学版),2017,43(2):144-149.
[14] Rosten E, Drummond T. Machine learning for high-speed corner detection[M]. Berlin Heidelberg: Springer, 2006.
[15] Calonder M, Lepetit V, Strecha C, et al. BRIEF: Binary Robust Independent Elementary Features[M]. Berlin Heidelberg: Springer, 2010.
[16] 孟凡清.基于背景差分法和ORB算法的运动目标检测与跟踪算法研究[D].北京:北京印刷学院,2014.
[17] 葛山峰,于莲芝,谢振.基于ORB特征的目标跟踪算法[J].电子科技,2017,30(2):98-100.
[18] 许宏科,秦严严,陈会茹.基于改进ORB的图像特征点匹配[J].科学技术与工程,2014,14(18):105-109.
[19] 王丽,郝晓丽.基于Kalman滤波器和改进Camshift算法的双眼跟踪[J].微电子学与计算机,2016,33(6):109-112.
备注/Memo
收稿日期: 2017-12-18
作者简介: 骆绍烨(1982—),男,讲师,研究方向为计算机视觉、信息检索.
基金项目: 国家自然科学基金资助项目(81373552); 莆田市科技计划项目(2015G2014); 莆田学院科研创新专项基金资助项目(2017024)