Facial action units (AUs) recognition is essential for emotion analysis and has been widely applied in mental state analysis. Existing work on AU recognition usually requires big face dataset with accurate AU labels. However, manual AU annotation requires expertise and can be time-consuming. In this work, we propose asemi-supervised approach forAUrecognition utilizing alargenumber of web face images without AU labels and a small face dataset with AU labels inspired by the co-training methods.

Neural Information Processing Systems http://nips.cc/

Several methods are proposed to keep on exploiting the correlation of different views. One straightforward way is completing the missing views,andthentheon-shelf multi-viewlearning algorithmscould beadopted. Themissing views are basically blockwise and thus low-rank based completion [12, 13] is not applicable which has been widely recognized [5, 14]. Missing modality imputation methods [15, 5] usually require samples with two paired modalities to train the networks which can predict the missing modality fromtheobservedone.

In practice, γ = 1 gives good results and we mainly tuneα,β. As mentioned in Sec 2.3, I(s;x1:T),I(z1:T;x1:T) are estimated contrastively.