Method for face modeling and related products
AssigneeGuangdong Oppo Mobile Telecommunications Corp., Ltd.
InventorQing YE
Life of this patent
Estimating the term…
Abstract
A method for face modeling and related products are provided. The method includes the following. Multiple first images of a face of an object at multiple different angles are obtained when the face rotates, where each first image is obtained at a different angle. Face detection is performed on any first image of the multiple first images, and the any first image is divided into a first face region and a first background region. Multiple depth images corresponding to the multiple first images are obtained and each of the multiple depth images is segmented according to the first face region and the first background region to obtain multiple face region images. Three-dimensional (3D) face modeling is performed according to the multiple face region images to obtain a 3D face image.
Claims20
A method for face modeling, comprising:
- obtaining a plurality of first images of a face of an object at a plurality of different angles when the face rotates, wherein each first image is obtained at a different angle;
- performing face detection on any first image of the plurality of first images, and dividing the any first image into a first face region and a first background region;
- obtaining a plurality of depth images corresponding to the plurality of first images, and segmenting each of the plurality of depth images according to the first face region and the first background region to obtain a plurality of face region images;
- and
- performing three-dimensional (3D) face modeling according to the plurality of face region images to obtain a 3D face image.
An electronic device, comprising:
- at least one camera configured to obtain a plurality of first images of a face of an object at a plurality of different angles when the face rotates, wherein each first image is obtained at a different angle;
- at least one processor coupled with the at least one camera;
- and
- a non-transitory computer readable storage, coupled to the at least one processor and storing at least one computer executable instruction thereon which, when executed by the at least one processor, causes the at least one processor to
- perform face detection on any first image of the plurality of first images, and divide the any first image into a first face region and a first background region;
- obtain a plurality of depth images corresponding to the plurality of first images, and segment each of the plurality of depth images according to the first face region and the first background region to obtain a plurality of face region images;
- and
- perform three-dimensional (3D) face modeling according to the plurality of face region images to obtain a 3D face image.
A non-transitory computer readable storage medium storing a computer program which, when executed by a processor, causes the processor to:
- obtain a plurality of first images of a face of an object at a plurality of different angles when the face rotates, wherein each first image is obtained at a different angle;
- perform face detection on any first image of the plurality of first images, and divide the any first image into a first face region and a first background region;
- obtain a plurality of depth images corresponding to the plurality of first images, and segment each of the plurality of depth images according to the first face region and the first background region to obtain a plurality of face region images;
- and
- perform three-dimensional (3D) face modeling according to the plurality of face region images to obtain a 3D face image.
Dependencies are inferred from the claim text.