行人检测技术综述_苏松志行人检测技术综述_苏松志.pdf

4 Vol .40 No. 4 2012 4 ACTA ELECTRONICA SINICA Apr. 2012 1, 2 1, 2 3 1, 2, 4 4 苏松志 , 李绍滋 , 陈淑媛 , 蔡国榕 , 吴云东 ( 1. , 361005; 2. , 361005; 3. , ; 4. , 361021) : , 2005-2011 . , , , , , . : ; ; ; : TP391 : A : 0372-2112 ( 2012) 04-0814-07 URL: http: / / www. ejourn l. org. cn DOI : 10. 3969/ j. issn. 0372-2112. 2012. 04. 031 A Survey on Pedestrian Detection 1, 2 1, 2 3 1, 2,4 4 SU Song-zhi , LI Sh o-zi , CHEN Shu-yu n , CAI Guo-rong ,WU Yun-dong ( 1. School of Inf ormation Science and Technology , Xiamen University , Xiamen, Fuj ian 361005 China; 2. Fuj ian Key aboratory of the Brain-like Intelligent Systems (Xiamen University ) , Xiamen, Fuj ian 361005, China; 3. Department of Comp uter Engineering and Science, Yuan-Ze University, Taiwan, China; 4. School of Science, Jimei University , Xiamen, Fuj ian 361021, China) Abstract : Pedestri n detection is n ctive re of rese rch with ch llenge in computer vision. This study conducts det iled survey on st te-of-the- rt pedestri n detection methods from 2005 to 2011, focusing on the two most import nt problems: fe ture ex- tr ction, the cl ssific tion nd loc liz tion. We divided these methods into different c tegories; pedestri n fe tures re divided into three subc tegories: low-level fe ture, le rning-b sed fe ture nd hybrid fe ture. On the other h nd, cl ssific tion nd loc liz tion is lso divided into two s