Client threshold prediction in biometric signature recognition by means of Multiple Linear Regression and its use for score normalization
作者：
Highlights：
• A new method for client threshold prediction in biometric signature is proposed.
• The proposal is based on Multiple Linear Regression, a well founded statistical tool.
• New robust parameters, not used before, have been successfully included in the model.
• The prediction model is optimized for each working point.
• The predicted threshold is used in score normalization improving the state of the art.
摘要
Highlights•A new method for client threshold prediction in biometric signature is proposed.•The proposal is based on Multiple Linear Regression, a well founded statistical tool.•New robust parameters, not used before, have been successfully included in the model.•The prediction model is optimized for each working point.•The predicted threshold is used in score normalization improving the state of the art.
论文关键词：Biometric signature recognition,Client threshold prediction,Score normalization,Multiple Linear Regression
论文评审过程：Received 15 January 2015, Revised 22 September 2015, Accepted 10 February 2016, Available online 23 February 2016, Version of Record 21 March 2016.
论文官网地址：https://doi.org/10.1016/j.patcog.2016.02.007
