Efficient 3D dental identification via signed feature histogram and learning keypoint detection
作者：
Highlights：
• We use a machine learning method to accurately detect keypoints on dental models.
• A novel local shape descriptor is proposed which can be efficiently computed.
• A highly efficient and robust dental identification algorithm is proposed.
• The proposed method is about 80 times faster than traditional 2D based methods.
• Promising results are achieved for single tooth identification.
摘要
Highlights•We use a machine learning method to accurately detect keypoints on dental models.•A novel local shape descriptor is proposed which can be efficiently computed.•A highly efficient and robust dental identification algorithm is proposed.•The proposed method is about 80 times faster than traditional 2D based methods.•Promising results are achieved for single tooth identification.
论文关键词：Dental biometrics,Tooth recognition,Postmortem identification,Shape descriptor,Keypoint detection,Shape matching,Random Forest
论文评审过程：Received 29 October 2014, Revised 11 September 2015, Accepted 10 May 2016, Available online 24 May 2016, Version of Record 3 June 2016.
论文官网地址：https://doi.org/10.1016/j.patcog.2016.05.007
