Fixed-Point Optimization Algorithm of AdaBoost Face Detection
-
摘要: 提出一种AdaBoost人脸检测的定点型优化算法,该算法以AdaBoost人脸检测原型算法为基础,分析了Cascade瀑布式级联分类器中弱分类器与强分类器分类计算的特点,有效分解了弱分类器与强分类器的计算过程,从而现实了强分类器与弱分类器相关模型参数有效分离标定。优化算法进一步利用图像积分图及弱分类器计算特点,完成对弱分类器计算过程及相关模型参数的定点型转化;同时,利用强分类器浮点的计算精度要求,完成强分类器计算过程及相关模型参数的定点型转化。该定点型AdaBoost人脸检测方法计算精度逼近原浮点型算法计算精度,保持了较高的人脸检测正确率,并利于后期的SIMD并行计算方法优化,同时,也利于算法在定点型嵌入式设备上的移植与优化。
-
[1] SUO Lu-jing, LU Xiao-feng, LU Heng-li, et al. Fast face detection method in real-time video[J]. Computer Engineering, 2011, 37(20): 166-168. [2] VIOLA P, MICHAEL J. Rapid object detection using a boosted cascade of simple features[C]//Proceedings of IEEE Conference on Computer Vision and Pattern Recognition. Kauai, Hawaii, USA: IEEE, 2001. [3] KEARNS M, VALIANT L G. Learning boolean formulae for finite automata is as hard as factoring[R]. Cambridge: Aiken Computation Laboratory, Harvard University, TR-1488, 1998. [4] LIU Xiao-ke, SUN Xie-hua, ZHOU Yong-xia. Multi-angle face detection based on new Haar-like feature[J]. Computer Engineering, 2009, 35(19): 195-197. [5] ABDEL-MOTTALEB M, ELGAMMAL A. Face detection in complex environments from color images[C]// Proceedings of IEEE Conference on Image Processing. [S.l.]: IEEE, 1999, 3: 622-626. [6] TANG Qi, SU Guang-da. Real time hardware face detection based on AdaBoost algorithm[J]. Computer Engineering, 2008, 34(7): 248-250. [7] KARLEKAR J, DESAI U B. Finding faces in color images using wavelet transform[C]//Proceedings of IEEE Conference on Image Analysis and Processing. Venice, Italy: IEEE, 1999: 1085-1088. [8] XU Jian-jun, ZHANG Rong, BI Du-yan, et al. An new AdaBoost video tracking algorithm[J]. Control and Decision, 2012, 27(5): 681-685. [9] ZHANG Yan-feng, HE Pei-kun. A revised AdaBoost algorithm—M-asy AdaBoost[J]. Transactions of Beijing Institute of Technology, 2011, 31(1): 64-68, 73. [10] ZHANG Z Q, ZHU L, LI S Z, et al. Real-time multi-view face detection[C]//Proceedings of The 5th International Conference on Automatic Face and Gesture Recognition. Washington, DC, USA: [s.n.], 2002. [11] 索璐静, 陆小锋, 陆亨立, 等. 实时视频中的快速人脸检测方法[J]. 计算机工程, 2011, 37(20): 166-168. [12] 刘晓克, 孙燮华, 周永霞. 基于新Haar-like特种的多角度人脸检测[J]. 计算机工程, 2009, 35(19): 195-197. [13] 唐奇, 苏光大. 基于AdaBoost算法的硬件实时人脸检测[J]. 计算机工程, 2008, 34(7): 248-250. [14] 徐建军, 张蓉, 毕笃彦, 等. 一种新的AdaBoost视频跟踪算法[J]. 控制与决策, 2012, 27(5): 681-685. [15] 张彦峰, 何佩琨. 一种改进的AdaBoost算法——M-asy AdaBoost[J]. 北京理工大学学报, 2011, 31(1): 64-68, 73. [16] 姚旭, 王晓丹, 张玉玺, 等. 基于AdaBoost和匹配追踪的选择性集成算法[J]. 控制与决策, 2014, 29(2): 208- 214. YAO Xu, WANG Xiao-dan, ZHANG Yu-xi, et al. Selective ensemble algorithm based on AdaBoost and matching pursuit[J]. Control and Decision, 2014, 29(2): 208-214.
点击查看大图
计量
- 文章访问数: 5075
- HTML全文浏览量: 129
- PDF下载量: 201
- 被引次数: 0