Computer Engineering and Applications ›› 2018, Vol. 54 ›› Issue (21): 188-193.DOI: 10.3778/j.issn.1002-8331.1707-0157

Previous Articles     Next Articles

Sketch face recognition combining with improved Surf feature

CAO Lin, LI Meng, ZHANG Hua   

  1. Department of Telecommunication Engineering, Beijing Information Science and Technology University, Beijing 100101, China
  • Online:2018-11-01 Published:2018-10-30

结合改进Surf特征的素描人脸识别

曹  林,李  猛,张  华   

  1. 北京信息科技大学 通信工程系,北京 100101

Abstract: The sketch face recognition belongs to the heterogeneous face recognition, and it is the research hot-spot in criminal investigation. According to the characteristics of sketch face recognition, the pseudo sketch transformation is used for the registered image, and the feature points are extracted by using Surf algorithm. Coordinate adjacent area consistency optimization have been proposed to eliminate the feature points of relative position inconsistent, and effective points of pseudo sketch image are calculated to achieve the purpose of recognition. The experiment are tested by using the existing sketch database, and the recognition rate is 99% when 50 feature points are selected. The result can prove that this algorithm is a new effective way in sketch face recognition. This method can be applied in sketch face recognition after optimization.

Key words: sketch face recognition, pseudo sketch, Surf algorithm, coordinate adjacent area consistency optimization

摘要: 素描人脸识别属于异质人脸识别范畴,是刑侦领域的研究热点。根据素描人脸识别的特点,对已配准的人脸图像进行伪素描转化,并用Surf算法提取伪素描图像对的特征点。对经过提取后的伪素描特征点进行坐标邻域一致性优化,排除坐标邻域相对位置不一致的特征点,最后统计伪素描图像对的有效特征点,以实现识别的目的。利用现有的素描人脸库,进行实验验证,在选取50个特征点时的识别率达到99%,验证了算法的有效性。该算法经优化后,可用于素描人脸识别。

关键词: 素描人脸识别, 伪素描, Surf算法, 坐标邻域优化