Spontaneous facial expression recognition by heterogeneous convolutional networks

Spontaneous facial expression achieves much attention recently as it has potential applications in the field of computer vision and pattern recognition. Although the convolutional networks have been applied for recognizing acted facial expressions and obtained the state-of-the-art performance, the performance of recognizing spontaneous facial expressions still needs to be improved. In this paper, a heterogeneous deep model is presented to recognize spontaneous expressions. The deep model consists of two types of convolutional networks with different architectures. To leverage the acted data, these two deep sub-networks are pre-trained over acted data and then transferred to the spontaneous data. Experiments have shown the advantages of the proposed method on the dataset of spontaneous facial expression.