Emotion Regression
Emotion regression, a subfield of affective computing, aims to predict the intensity of emotions rather than simply classifying them into discrete categories. Current research focuses on developing robust models, including those based on temporal convolutional networks (TCNs) and incorporating multiple modalities like audio and video, to improve accuracy and handle missing data. This work is significant for advancing human-computer interaction, enabling the creation of more empathetic and responsive systems in areas such as assistive technologies and mental health monitoring, while also highlighting the importance of addressing biases in these systems.
Papers
February 16, 2024
August 30, 2023
May 12, 2023
April 11, 2023
April 7, 2022