Feature selection for face recognition based on multi-objective evolutionary wrappers

作者:

Highlights:

• We describe a methodology for the selection of relevant features for face recognition.

• Multiobjective wrappers are designed by means of genetic algorithms.

• We aim to minimize the number of features while maximizing discriminability.

• Results show that the proposed approach provides improved classification performance.

摘要

•We describe a methodology for the selection of relevant features for face recognition.•Multiobjective wrappers are designed by means of genetic algorithms.•We aim to minimize the number of features while maximizing discriminability.•Results show that the proposed approach provides improved classification performance.

论文关键词:Wrappers,Multi-objective genetic algorithms,Feature selection,Face recognition

论文评审过程:Available online 21 March 2013.

论文官网地址:https://doi.org/10.1016/j.eswa.2013.03.032