Face Matching with Hybrid Features

摘要：

This paper addresses the effectiveness of utilizing hybrid features for face identification. The proposed method uses facial features that are represented by feature vectors derived from Gabor wavelets and Zernike moments, which provide both local and global image descriptors, respectively. We demonstrate empirically the robustness of combining these global and local feature descriptors. The contribution of Zernike moments in rectifying small rotations of features, which may occur due to changes in facial expression, is investigated. Experimental results indicate that the utilization of the aforementioned hybrid features yeilds more accurate recognition rates compared to using a single feature type alone.

作者: Menaka Rajapakse

作者单位: RWCP, Multi-Modal Functions KRDL 21 Heng Mui Keng Terrace, Singapore 119613

会议类型: 国际会议

会议名称: 8th International Conference on Neural Information Processing(ICONIP 2001)(第八届国际神经信息处理大会)

会议地点: 上海

会议语种:英文

页码: 218-221

在线出版日期: 2001-11-14（万方平台首次上网日期，不代表论文的发表时间）