SDTF-Net: Static and dynamic time-frequency network for Speech Emotion Recognition

Lu-Yao Liu, Wenzhe Liu, Lin Feng

Abstract: Highlights•A new framwork to comprehensively extract speech emotional features.•TF-CNN module to capture static emotional features based on TF characteristics.•TF-Trans module to capture dynamic emotional features based on TF characteristics.•SD-Cross module to effectively integrate the static and dynamic emotional features.