A further study of low resolution androgenic hair patterns as a soft biometric trait
作者：
Highlights：
• Low resolution androgenic hair patterns contain rich distinctive information.
• Large feature sets (about 60,000 features) are employed for leg identification.
• Class specific partial least squares models are obtained from large feature sets.
• Multi-directional gridding systems increase discriminative power and robustness.
• The positive sample generation scheme helps tackle viewpoint and pose variations.
摘要
•Low resolution androgenic hair patterns contain rich distinctive information.•Large feature sets (about 60,000 features) are employed for leg identification.•Class specific partial least squares models are obtained from large feature sets.•Multi-directional gridding systems increase discriminative power and robustness.•The positive sample generation scheme helps tackle viewpoint and pose variations.
论文关键词：Soft biometrics,Biometrics,Emerging biometrics,Forensics
论文评审过程：Received 2 May 2016, Revised 25 March 2017, Accepted 28 August 2017, Available online 8 September 2017, Version of Record 6 February 2018.
论文官网地址：https://doi.org/10.1016/j.imavis.2017.08.009
