Provided are a method and a system for recognizing a complex multi-emotion of a user face. The complex multi-emotion recognizing method comprises the steps of: searching a face area of a user in an input image; generating an average shape model; fitting the average shape model to a user face; finding a mesh based on a preset reference point; finding an object variation in an area of interest limited by the mesh; measuring similarity based on the object variation; and determining strength of a plurality of emotions according to a frequency among emotion expressions preset inside a database by the similarity.
展开▼