A Curvelet-based approach for textured 3D face recognition
作者：
Highlights：
• Identifying distinctive keypoints on textured 3D face surfaces rich with features.
• These keypoints are identified in the Curvelet domain across mid-frequency bands.
• The repeatability of these keypoints is high in both neutral and nonneutral faces.
• Building local surface descriptors around the keypoints in the Curvelet domain.
• Reported results show superior performance on three datasets, namely FRGC, BU-3DFE and Bosphorus, compared to prior art.
摘要
Highlights•Identifying distinctive keypoints on textured 3D face surfaces rich with features.•These keypoints are identified in the Curvelet domain across mid-frequency bands.•The repeatability of these keypoints is high in both neutral and nonneutral faces.•Building local surface descriptors around the keypoints in the Curvelet domain.•Reported results show superior performance on three datasets, namely FRGC, BU-3DFE and Bosphorus, compared to prior art.
论文关键词：Face recognition,Keypoint detection,Local features,Digital Curvelet transform
论文评审过程：Received 23 February 2014, Revised 7 June 2014, Accepted 11 October 2014, Available online 23 October 2014.
论文官网地址：https://doi.org/10.1016/j.patcog.2014.10.013
