Cross-Sensor Evaluation of Textural Descriptors for Gender Prediction from Fingerprints

Abstract

Estimating gender from fingerprints brings benefits to various security, forensic and intelligence applications. However, achieving high prediction accuracy without human intervention is currently a challenge. Furthermore, biometric data may be originated from different sensors; thus, analyzing the sensitivity of the feature set to acquisition device changes becomes important. This paper evaluates performance of three local textural descriptors combined with image quality and minutiae count for automatic gender estimation from fingerprint images acquired using four different optical sensors and TenPrint cards. In particular, Local Binary Patterns (LBP), Local Phase Quantization (LPQ) and Binarized Statistical Image Features (BSIF) features were concatenated with image quality NFIQ2 and minutiae count. Such a study explores robustness and degradation of these features with respect to capture bias. Additionally, logistic regression models are applied to identify the significant features for gender estimation.

Cite

Text

Marasco et al. "Cross-Sensor Evaluation of Textural Descriptors for Gender Prediction from Fingerprints." IEEE/CVF Winter Conference on Applications of Computer Vision Workshops, 2019. doi:10.1109/WACVW.2019.00017

Markdown

[Marasco et al. "Cross-Sensor Evaluation of Textural Descriptors for Gender Prediction from Fingerprints." IEEE/CVF Winter Conference on Applications of Computer Vision Workshops, 2019.](https://mlanthology.org/wacvw/2019/marasco2019wacvw-crosssensor/) doi:10.1109/WACVW.2019.00017

BibTeX

@inproceedings{marasco2019wacvw-crosssensor,
  title     = {{Cross-Sensor Evaluation of Textural Descriptors for Gender Prediction from Fingerprints}},
  author    = {Marasco, Emanuela and Cando, Stefany and Tang, Larry and Tabassi, Elham},
  booktitle = {IEEE/CVF Winter Conference on Applications of Computer Vision Workshops},
  year      = {2019},
  pages     = {55-62},
  doi       = {10.1109/WACVW.2019.00017},
  url       = {https://mlanthology.org/wacvw/2019/marasco2019wacvw-crosssensor/}
}