JISP  >> Vol. 5 No. 4 (October 2016)

    基于颜色空间纹理融合特征的SOAMST人脸跟踪算法
    SOAMST Face Tracking Algorithm Based on the Combined Feature of Color Space and Texture

  • 全文下载: PDF(826KB) HTML   XML   PP.166-173   DOI: 10.12677/JISP.2016.54019  
  • 下载量: 949  浏览量: 2,203  

作者:  

周真真,冯子亮:四川大学计算机学院,成都 四川;视觉合成图形图像技术国家重点学科实验室,成都 四川

关键词:
人脸跟踪SOAMST二阶空间直方图MBLBP颜色直方图Face Tracking SOAMST The Second-Order Spatial Histogram MBLBP Color Histogram

摘要:

针对人脸跟踪中存在的人脸目标与背景颜色相似、光照变化对跟踪算法鲁棒性的影响,本文提出基于颜色空间纹理融合特征的SOAMST跟踪算法,将二阶空间直方图代替SOAMST建立目标模型中的颜色直方图,二阶直方图融合了目标颜色信息和颜色的空间分布信息,比传统颜色直方图更具有目标鉴别能力。此外与MBLBP特征进行融合,构建有效的联合直方图建立目标模型。不同场景下的实验结果表明,与原算法比较,本文的改进算法在人脸颜色与背景相似,光照变化,旋转大小变化等复杂环境下,抗干扰性更强,有更好的跟踪效果,具有更高的准确性和鲁棒性。

In view of the existing question that in the human face tracking, face target is similar to the back-ground color, and light changes impact on the robustness of tracking algorithm, in this paper we put forward a SOAMST face tracking algorithm based on the combined feature of color space and texture, which chooses the second-order spatial histogram instead of SOAMST color histogram to build target model. The second-order histogram is combined with color’s information and color spatial distribution’s information, compared to the traditional color histogram with more target identification ability. In addition to the fusion of MBLBP characteristics, it can build effective joint histogram to establish target model. Experimental results in different scenarios show that compared with the original algorithm, the improved algorithm has better anti-interference performance, better tracking effect, and a higher accuracy and robustness under the complex environment which face color is similar to the background, light changes, and the size of the changes.

文章引用:
周真真, 冯子亮. 基于颜色空间纹理融合特征的SOAMST人脸跟踪算法[J]. 图像与信号处理, 2016, 5(4): 166-173. http://dx.doi.org/10.12677/JISP.2016.54019

参考文献

[1] Comaniciu, D., Ramesh, V. and Meer, P. (2000) Real-Time Tracking of Using Nonrigid Objects Using Mean Shift. IEEE Conference on Computer Vision and Pattern Recognition, 15-15 June 2000, 2142.
[2] Gordon, N., Salmond, D. and Smith, A. (2002) Novel Approach to Nonlinear and Non-Gaussian Bayesian State Estimation. IEE Proceedings F, 140, 107-113.
[3] Ning, J., Zhang, L., Zhang, D., et al. (2012) Scale and Orientation Adaptive Mean Shift Tracking. IET Computer Vision, 6, 52-61. http://dx.doi.org/10.1049/iet-cvi.2010.0112
[4] 侯建华, 黄奇, 郑桂林, 等. 一种尺度和方向适应性的Mean Shift跟踪算法[J]. 中南民族大学学报(自然科学版), 2015, 34(1): 83-88.
[5] 朱志宇. 粒子滤波算法及其应用[M]. 北京: 科学出版社, 2010: 194-195.
[6] Birchfield, S. and Rangarajan, S. (2005) Spatiograms versus Histograms for Region-Based Tracking. Proceedings of the IEEE Conference on Computer Vision and Pattern Recognition, San Diego, 20-25 June 2005, 1158-1163. http://dx.doi.org/10.1109/cvpr.2005.330
[7] Ojala, T., Pietikäinen, M. and Harwood, D. (1994) Performance Evaluation of Texture Measures with Classification Based on Kullback Discrimination of Distributions. Proceedings of the 12th IAPR International Conference on Pattern Recognition (ICPR 1994), 1, 582-585. http://dx.doi.org/10.1109/ICPR.1994.576366
[8] Ojala, T., Pietikäinen, M. and Harwood, D. (1976) A Comparative Study of Texture Measures with Classification Based on Feature Distributions. Pattern Recognition, 29, 51-59. http://dx.doi.org/10.1016/0031-3203(95)00067-4
[9] Liao, S.C., Zhu, X.X., Lei, Z., Zhang, L. and Li, S.Z. (2007) Learning Multi-Scale Block Local Binary Patterns for Face Recognition. International Conference on Biometrics (ICB), 828-837. http://dx.doi.org/10.1007/978-3-540-74549-5_87
[10] 刘明宝, 姚鸿勋, 高文. 彩色图像的实时人脸跟踪方法[J]. 计算机学报, 1998, 21(6): 527-532.
[11] 李菊, 余烨. 基于颜色和LBP多特征的mean shift的跟踪算法[J]. 合肥工业大学学报(自然科学版), 2014, 5(37): 578-581.
[12] 王保云, 范保杰. 基于颜色纹理联合特征直方图的自适应Meanshift跟踪算法[J]. 南京邮电大学学报(自然科学版), 2013(3): 18-25.