A facial expression-based consecutive emotion recognition method of a robot, comprises: a step of constructing a database to store images of each emotion; a step of calculating an independent variable through a principal component analysis (PCA) based on distribution of images stored in the database; a step of learning a linear regression equation using the calculated independent variable; and a step of applying an input image to the linear regression equation to calculate an emotional state value. Accordingly, a robot continuously estimates human emotions through human facial expressions such that emotions are able to be shared between the robot and humans.
展开▼
