Action Units That Constitute Trainable Micro-expressions (and A Large-scale Synthetic Dataset). (arXiv:2112.01730v4 [cs.CV] UPDATED)

Web: http://arxiv.org/abs/2112.01730

Because of the expensive data collection process, micro-expression (MiE)
datasets are generally much smaller in scale than those in other computer
vision fields, rendering large-scale training less feasible. This paper
develops a protocol to automatically synthesize MiE training data that 1) are
of a large scale and 2) allow us to train accurate recognition models for
real-world test data. Specifically, we discover three types of Action Units
(AUs) that can constitute trainable MiEs. These AUs come from real-world MiEs,
early …

arxiv cv dataset scale