A bullet screen song requesting method and
system based on face identification are disclosed. The method includes the steps of S1 setting a bullet screen container for containing bullet screens, storing rendered dynamic images in advance, dividing the rendered dynamic images into different application scenes, and setting song-requesting bullet screen
asterisk wildcards, S2 obtaining input character information of a user, obtaining
video image information through a camera, and determining whether the face of a user is contained in the
video image information, S3 conducting dynamic tracking for the detected user face, conducting target searching for the dynamic tracked user face in a facial
image database to realize face identification, detecting key points of a face, and determining the
mood of the user according to the key points, S4 selecting corresponding application scene according to the
mood detected in the S3, obtaining the rendered dynamic image from the application scene and superposing the rendered dynamic image onto the key points of the face, setting the display time length, entering position and moving track on the bullet screen, and returning to S2 until video communication ends.