Method and apparatus for facial recognition

• US 10,691,928 B2
• Filed: 07/31/2018
• Issued: 06/23/2020
• Est. Priority Date: 09/21/2017
• Status: Active Grant

• Current Assignee: Baidu Online Network Technology (Beijing) Co., Ltd (Baidu Incorporated)
• Original Assignee: Baidu Online Network Technology (Beijing) Co., Ltd (Baidu Incorporated)
• Inventors: Du, Kang
• Primary Examiner(s): Alam, Fayyaz
• Application Number:
  US16/050,381
• Publication Number:
  US 20190087647A1
• Time in Patent Office: 693 Days
• Field of Search: None
• CPC Class Codes: G06N 3/045 Combinations of networks
• CPC Class Codes: G06N 3/08 Learning methods
• CPC Class Codes: G06V 40/169 Holistic features and representations, i.e. based on the facial image taken as a whole
• CPC Class Codes: G06V 40/171 Local features and components; Facial parts eye characteristics G06V40/18; Occluding parts, e.g. glasses; Geometrical relationships
• CPC Class Codes: G06V 40/172 Classification, e.g. identification

Abstract

Embodiments of the present disclosure disclose a method and apparatus for facial recognition. A specific embodiment of the method includes: acquiring a to-be-recognized image; inputting the to-be-recognized image into a pre-trained first convolutional neural network to obtain complete facial feature information and partial facial feature information, the first convolutional neural network being used to extract a complete facial feature and a partial facial feature; and inputting the complete facial feature information and the partial facial feature information into a pre-trained second convolutional neural network to obtain a facial recognition result, the second convolutional neural network being used to represent a correlation between the facial recognition result, and the complete facial feature information and the partial facial feature information. This embodiment improves the accuracy of the recognition result in a situation where a face is partially covered.

Claims

1. A method for recognizing a face, comprising:

acquiring a to-be-recognized image;

inputting the to-be-recognized image into a pre-trained first convolutional neural network to obtain complete facial feature information and partial facial feature information, the first convolutional neural network being used to extract a complete facial feature and a partial facial feature; and

inputting the complete facial feature information and the partial facial feature information into a pre-trained second convolutional neural network to obtain a facial recognition result, the second convolutional neural network being used to represent a correlation between the facial recognition result, and the complete facial feature information and the partial facial feature information,

wherein the method is performed by at least one processor.

2. The method according to claim 1, wherein the complete facial feature information comprises a plurality of complete facial feature maps, and the plurality of complete facial feature maps comprise a first complete facial feature map and a plurality of second complete facial feature maps, each point in the first complete facial feature map represents a confidence level indicating an existence of a complete face in an area in the to-be-recognized image corresponding to the each point, and each point in each of the plurality of second complete facial feature maps represents position information of an area in the to-be-recognized image corresponding to the point in the each of the plurality of second complete facial feature maps.

3. The method according to claim 2, wherein the plurality of second complete facial feature maps comprise four second complete facial feature maps, and points in the four second complete facial feature maps respectively represent a horizontal coordinate of an upper-left vertex, a vertical coordinate of the upper-left vertex, a horizontal coordinate of a lower-right vertex, and a vertical coordinate of the lower-right vertex of a corresponding area in the to-be-recognized image.

4. The method according to claim 1, wherein the partial facial feature information comprises a plurality of partial facial feature maps, and the plurality of partial facial feature maps comprise at least one first partial facial feature map, and a plurality of second partial facial feature maps corresponding to each of the at least one first partial facial feature map, wherein each point in the each of the at least one first partial facial feature map represents a confidence level indicating an existence of a partial face in an area in the to-be-recognized image corresponding to the point in the each of the first partial facial feature map, and each point in each of the plurality of second partial facial feature maps represents position information of an area in the to-be-recognized image corresponding to the point in the each of the plurality of second partial facial feature maps.

5. The method according to claim 4, wherein the partial face comprises at least one of:

an eye, a nose, or a mouth.

6. The method according to claim 4, wherein the plurality of second partial facial feature maps corresponding to the each of the at least one first partial facial feature map comprise four second partial facial feature maps, and points in the four second partial facial feature maps respectively represent a horizontal coordinate of an upper-left vertex, a vertical coordinate of the upper-left vertex, a horizontal coordinate of a lower-right vertex of the area, and a vertical coordinate of the lower-right vertex of a corresponding area in the to-be-recognized image.

7. The method according to claim 1, wherein the second convolutional neural network is a fully convolutional network, and a number of convolutional kernels in a last convolutional layer of the fully convolutional network is five.

8. An apparatus for facial recognition, comprising:

at least one processor; and

a memory storing instructions, the instructions when executed by the at least one processor, cause the at least one processor to perform operations, the operations comprising;

acquiring a to-be-recognized image;

inputting the to-be-recognized image into a pre-trained first convolutional neural network to obtain complete facial feature information and partial facial feature information, the first convolutional neural network being used to extract a complete facial feature and a partial facial feature; and

inputting the complete facial feature information and the partial facial feature information into a pre-trained second convolutional neural network to obtain a facial recognition result, the second convolutional neural network being used to represent a correlation between the facial recognition result, and the complete facial feature information and the partial facial feature information.

9. The apparatus according to claim 8, wherein the complete facial feature information comprises a plurality of complete facial feature maps, and the plurality of complete facial feature maps comprise a first complete facial feature map and a plurality of second complete facial feature maps, each point in the first complete facial feature map represents a confidence level indicating an existence of a complete face in an area in the to-be-recognized image corresponding to the each point, and each point in each of the plurality of second complete facial feature maps represents position information of an area in the to-be-recognized image corresponding to the point in the each of the plurality of second complete facial feature maps.

10. The apparatus according to claim 9, wherein the plurality of second complete facial feature maps comprise four second complete facial feature maps, and points in the four second complete facial feature maps respectively represent a horizontal coordinate of an upper-left vertex, a vertical coordinate of the upper-left vertex, a horizontal coordinate of a lower-right vertex, and a vertical coordinate of the lower-right vertex of a corresponding area in the to-be-recognized image.

11. The apparatus according to claim 8, the partial facial feature information comprises a plurality of partial facial feature maps, and the plurality of partial facial feature maps comprise at least one first partial facial feature map, and a plurality of second partial facial feature maps corresponding to each of the at least one first partial facial feature map, wherein each point in the each of the at least one first partial facial feature map represents a confidence level indicating an existence of a partial face in an area in the to-be-recognized image corresponding to the point in the each of the first partial facial feature map, and each point in each of the plurality of second partial facial feature maps represents position information of an area in the to-be-recognized image corresponding to the point in the each of the plurality of second partial facial feature maps.

12. The apparatus according to claim 11, wherein the partial face comprises at least one of:

an eye, a nose, or a mouth.

13. The apparatus according to claim 11, wherein the plurality of second partial facial feature maps corresponding to the each of the at least one first partial facial feature map comprise four second partial facial feature maps, and points in the four second partial facial feature maps respectively represent a horizontal coordinate of an upper-left vertex, a vertical coordinate of the upper-left vertex, a horizontal coordinate of a lower-right vertex of the area, and a vertical coordinate of the lower-right vertex of a corresponding area in the to-be-recognized image.

14. The apparatus according to claim 8, wherein the second convolutional neural network is a fully convolutional network, and a number of convolutional kernels in a last convolutional layer of the fully convolutional network is five.

15. A non-transitory computer-readable storage medium storing a computer program, the computer program when executed by one or more processors, causes the one or more processors to perform operations, the operations comprising:

acquiring a to-be-recognized image;

inputting the to-be-recognized image into a pre-trained first convolutional neural network to obtain complete facial feature information and partial facial feature information, the first convolutional neural network being used to extract a complete facial feature and a partial facial feature; and

inputting the complete facial feature information and the partial facial feature information into a pre-trained second convolutional neural network to obtain a facial recognition result, the second convolutional neural network being used to represent a correlation between the facial recognition result, and the complete facial feature information and the partial facial feature information.