A method may include playing a first song on a device, capturing an image of a user, performing facial expression recognition of the user based on the image, and selecting a second song based on a facial expression of the user.