A novel 2D and 3D multimodal approach for in-the-wild facial expression recognition

作者：

Highlights：

• Reconstruction of 3D face data from existing 2D face images is useful for in-the-wild facial expression recognition (FER).

• A novel and efficient deep learning approach with the fusion of 2D and 3D modalities in in-the-wild FER

• Extensive experiments on learning 3D point cloud data for FER

• Achieving state-of-the-art performances on RAF, SFEW, AffectNet, BU-3DFE dataset

摘要

•Reconstruction of 3D face data from existing 2D face images is useful for in-the-wild facial expression recognition (FER).•A novel and efficient deep learning approach with the fusion of 2D and 3D modalities in in-the-wild FER•Extensive experiments on learning 3D point cloud data for FER•Achieving state-of-the-art performances on RAF, SFEW, AffectNet, BU-3DFE dataset

论文关键词：Facial expression recognition,In-the-wild FER,2D + 3D FER

论文评审过程：Received 15 April 2019, Revised 27 September 2019, Accepted 16 October 2019, Available online 28 October 2019, Version of Record 1 November 2019.

论文官网地址：https://doi.org/10.1016/j.imavis.2019.10.003