| 摘要 |
PURPOSE: A method and a device for recognizing a face using a 3-D(Dimensional) face descriptor are provided to recognize a face photo, the video including the face, and a 3-D face mesh model by using the 3-D face descriptor generated from the image having many poses, a video stream having face rotation, and the 3-D face mesh model. CONSTITUTION: A basis generator(11) generates/stores an entire feature space basis including a pose value in a database by registering/training one of the images having many poses, the video stream having the face rotation, and the 3-D face mesh model. An entire face descriptor generator(13) generates/stores the 3-D entire face descriptor in the database by using one of the images, the video stream, and the 3-D face mesh model, and the pose value provided from the basis generator. A partial face descriptor generator(15) generates a 3-D partial face descriptor by using one of the images, the video stream, and the 3-D face mesh model, and the pose value, and generates a 3-D partial face descriptor database by using the pose value and the 3-D partial face descriptor. A searcher(17) measures similarity by searching the 3-D partial face descriptor database and outputs a search result.
|
