We propose a time-aware sampling network (TAS-Net) using deep reinforcement learning (DRL) for unsupervised emotion recognition, which is able to detect key emotion moments and disregard irrelevant and misleading parts.
Thanks to the MOEKLINNS Lab, Department of Computer Science and Technology, Xi’an Jiaotong University, which provided the labels of key emotion fragments in the SEED database. For details, please refer to the work of Zhang et al.