Facial Expression Recognition with Manifold Learning and Graph Convolutional Network

Visibility Convolutional Neural Networks for Label Distribution Learning of Facial Emotions

“…Apart from the individual's emotional state, such expressions convey valuable information regarding their identity and cultural traits. This information has been widely exploited by several methods to describe human sentiments, such as in the form of the two-dimensional (2D) valencearousal coordinate space [1], [2], [3], and via manifold learning of facial expressions [4]. In contrast to these methods, there are categorical approaches that map a facial expression to one or more emotions.…”