期刊文献+

梯度方向直方图特征的人耳身份识别方法 被引量:3

Ear recognition with histogram of oriented gradient features
下载PDF
导出
摘要 人耳特征体具有普遍性、唯一性、稳定性和可采集性等作为生物特征识别必备的基本特性,易实现非打扰式识别.研究一类新型的基于梯度方向直方图的人耳身份识别方法,将人耳图像划分为不同子区域,分别提取各子区域梯度方向直方图特征,采用模糊隶属度匹配融合策略获取识别结果.重点讨论了局部遮挡时基于梯度方向直方图特征描述的人耳识别问题,定义并分析了人耳子区域识别贡献率,探讨累计遮挡问题,可为人耳图像有效鉴别子区域的确定提供参考.实验结果表明,基于梯度方向直方图的人耳身份识别方法是可行的与有效的. Ear biometric has the basic characteristics, such as universality, distinctiveness, stability and collectibility. The non-disturbed recognition can be easily realized by ear biometric or fusion of ear and face biometrics. As a novel biometric identification, ear recognition is a very active research project in the field of pattern recognition. However, ear recognition result is usually affected by the head posture of image, illumination variations, occlusion and other factors, which makes the process of recognition more complicated and challenging. In this paper, a novel ear recognition approach based on histogram of oriented gradient feature extraction is studied. The ear recognition scheme by combining histogram of oriented gradient features with sub-region fuzzy fusion is presented. The ear image is divided into a number of sub-regions, histograms of oriented gradient features of different sub-regions are extracted separately, and the fuzzy membership matching fusion strategy is introduced to obtain the last classification label. Furthmore, the issue on ear recognition under local occlusion based on histogram of oriented gradient is discussed in detail. The recognition contribution rate for each sub-region is defined and analyzed, and the accumulated occlusion is investigated, which can give a reference for determining the effectiveidentify regional block. The maximum occluded range and effective identify regional block can effentively identify the human ear images. There{ore, the standard of the ear images whether retained or discarded can be determined. Experimental results show that ear recognition approach based on histogram of oriented gradient features is practicable and effective.
出处 《南京大学学报(自然科学版)》 CAS CSCD 北大核心 2012年第4期452-458,共7页 Journal of Nanjing University(Natural Science)
基金 国家自然科学基金(60573058) 河北省科技支撑计划重点项目(10213516D) 河北省自然科学基金(F2010001106)
关键词 人耳识别 梯度方向直方图 模糊融合匹配 子区域 局部遮挡 ear biometric recognition, histogram of oriented gradient, fuzzy fusion match, sub-region, local occlusion
  • 引文网络
  • 相关文献

参考文献16

二级参考文献106

共引文献89

同被引文献38

  • 1袁立,穆志纯,徐正光,刘克.基于人耳生物特征的身份识别[J].模式识别与人工智能,2005,18(3):310-315. 被引量:25
  • 2汪沁,江淑红,张建秋,胡波.提高Mean-shift跟踪算法性能的方法[J].复旦学报(自然科学版),2007,46(1):85-90. 被引量:12
  • 3WU Yiming, CHAN KAP-LUK. An extended Isomap algorithm for learning multi-class manifold [C]// Proceedings of 2004 International Conference on Machine Learning and Cybernetics, Shanghai, China, August 26-29, 2004, 6: 3429-3433.
  • 4Amjad Ali. Fingerprint Preprocessing and Texture-based Feature Extraction Algorithms for Personal Recognition [D]. 北京:北京邮电大学,2012:1-5.
  • 5Victor B, Bowyer K, Sarkar S. An evaluation of face and ear biometrics [C]// Proceedings 16th International Conference on Pattern Recognition, Florida, USA, 2002: 429-432.
  • 6Zhang D Q, Zhou Z H, Chen S C. Diagonal principal component analysis for face recognition [J], Pattern Recognition(S0031-3203), 2006, 39(1): 140-142.
  • 7Zhang X, Ren X. Two Dimensional Principal Component Analysis based Independent Component Analysis for face recognition [C]//2011 International Conference on Multimedia Technology, Hangzhou, china, July 26-28, 2011 : 934-936.
  • 8Liao P, Liu J, Wang M, et al. Ensemble local fractional LDA for face recognition [C]// Computer Science and Automation Engineering, Zhangjiajie, China, May25-27, 2012: 586-590.
  • 9Seung H S, Lee D D. The Manifold Ways of Perception [J]. Science(S0036-8075), 2000, 290(5500): 2268-2269.
  • 10Vural E, Frossard P. Learning pattern transformation manifolds for classification [C]//Image Processing, Orlando, FL, Sept. 30-Oct. 3, 2012: 1165-1168.

引证文献3

二级引证文献16

;
使用帮助 返回顶部