Comparing ensemble strategies for deep learning: An application to facial expression recognition

作者：

Highlights：

• CNNs can be used as solutions to the Facial Expression Recognition (FER) problem.

• Different CNN ensembling strategies are experimentally compared in the FER context.

• Exploiting different sources of variability is crucial to improve the accuracy of a CNN ensemble.

• Pre-training is found to be the most effective strategy to build CNNs ensembles for FER.

摘要

•CNNs can be used as solutions to the Facial Expression Recognition (FER) problem.•Different CNN ensembling strategies are experimentally compared in the FER context.•Exploiting different sources of variability is crucial to improve the accuracy of a CNN ensemble.•Pre-training is found to be the most effective strategy to build CNNs ensembles for FER.

论文关键词：Facial Expression Recognition,Convolutional Neural Networks,Ensemble learning,Ensemble construction

论文评审过程：Received 6 September 2018, Revised 18 May 2019, Accepted 13 June 2019, Available online 14 June 2019, Version of Record 20 June 2019.

论文官网地址：https://doi.org/10.1016/j.eswa.2019.06.025