Multi-view dynamic facial action unit detection

作者：

Highlights：

• Hollistic Action Unit detectors that are based on convolutional neural networks

• Merging RGB and OF information in order to boost AU classification (AUNets)

• Extending AUNets to cope with three-dimensional rendered facial images

• State-of-the-art in BP4D dataset and FERA17 Challenge

摘要

•Hollistic Action Unit detectors that are based on convolutional neural networks•Merging RGB and OF information in order to boost AU classification (AUNets)•Extending AUNets to cope with three-dimensional rendered facial images•State-of-the-art in BP4D dataset and FERA17 Challenge

论文关键词：

论文评审过程：Received 1 August 2017, Revised 21 May 2018, Accepted 14 September 2018, Available online 26 September 2018, Version of Record 13 May 2022.

论文官网地址：https://doi.org/10.1016/j.imavis.2018.09.014