改进PSO与K均值聚类肤色分割的人脸检测算法  

Face detection method with improved PSO and K-means clustering based on skin color segmentation

作　　者：班俊硕 赖惠成[1] 林宪峰[1] 杨敏[1] 董九玲 

机构地区：[1]新疆大学信息科学与工程学院,乌鲁木齐830046

出　　处：《激光杂志》2017年第2期82-86,共5页Laser Journal

基　　金：国家自然科学基金资助项目(61561048);新疆维吾尔自治区科学基金资助项目(2015211C257)

摘　　要：为了提高在强光照等复杂环境下人脸图像检测的鲁棒性和准确率。提出在YCgCr彩色空间,通过改进的粒子群(Particle Swarm Optimization,PSO)与K均值聚类综合的方法进行肤色分割,提升了聚类方法的全局检索能力;再对分割后的肤色区域进行二值形态学和人脸几何形状特征处理,去除人脸区域以外噪声,得出候选人脸区;最终通过改进AdaBoost算法对候选人脸区域进行检测验证。仿真实验表明,该算法人脸检测正确率高,鲁棒性和适应性好,具有很强的运用价值。To improve the strong light, complex environment Human face image detection robustness and accuracy rate. Presented at YCgCr color space, through improved particle swarm optimization （PSO） were integrated with the K -means clustering segmentation method of color to enhance the global search capabilities of clustering method; and then the color area of the divided binary morphological and people face geometry features, thereby removing the noise outside face region, drawn face region candidate; final verification through improved AdaBoost algorithm to detect face candidate region. Simulation results show that the proposed algorithm face detection correct rate, robustness and adapt- ability, a strong application value.

关 键 词：YCGCR颜色空间 K均值算法 粒子群算法 二值形态学 几何特征 改进AdaBoost 

分 类 号：TN911[电子电信—通信与信息系统]