Face verification method and apparatus, server and readable storage medium

1. A face verification method, comprising:

acquiring a two-dimensional face image and a three-dimensional face image for face verification of a target user;

performing face recognition on the acquired two-dimensional face image with respect to the target user to obtain a result of the face recognition;

in response to the result of the face recognition indicating that the face recognition is successful, performing three-dimensional reconstruction on the acquired two-dimensional face image to obtain a reconstructed three-dimensional face image;

performing liveness detection on the acquired three-dimensional face image to determine whether a face in the acquired three-dimensional face image corresponds to a live person;

in response to determining that the face in the acquired three-dimensional face image corresponds to a live person, comparing the reconstructed three-dimensional face image with the acquired three-dimensional face image to determine a similarity value; and

in response to the determined similarity value satisfying a condition, determining that the two-dimensional face image corresponds to the target user.

2. The method of claim 1 , wherein the comparing the reconstructed three-dimensional face image with the acquired three-dimensional face image to determine a similarity value comprises:

comparing three-dimensional structure data of the reconstructed three-dimensional face image with three-dimensional structure data of the acquired three-dimensional face image.

3. The method of claim 1 , wherein the condition comprises:

that the determined similarity value is not less than a threshold similarity value.

4. The method of claim 1 , wherein the performing face recognition on the acquired two-dimensional face image with respect to the target user to obtain a result of the face recognition comprises:

performing face recognition on the two-dimensional face image to obtain a face recognition value;

determining whether the face recognition value is not less than a threshold value;

obtaining a result indicating that the face recognition is successful if the face recognition value is not less than the threshold value; and

obtaining a result indicating that the face recognition is unsuccessful if the face recognition value is less than the threshold value.

5. The method of claim 1 , wherein the performing liveness detection on the acquired three-dimensional face image to determine whether a face in the acquired three-dimensional face image corresponds to a live person comprises:

performing liveness detection on the acquired three-dimensional face image to obtain a liveness detection value;

determining whether the liveness detection value is less than a threshold liveness value;

if the liveness detection value is not less than the threshold liveness value, determining that the face in the acquired three-dimensional face image corresponds to a live person; and

if the liveness detection value is less than the threshold liveness value, determining that the face in the acquired three-dimensional face image does not correspond to a live person.

6. The method of claim 1 , further comprising, before performing face recognition on the acquired two-dimensional face image:

subtracting an average brightness value from the acquired two-dimensional face image; or

performing face detection, face alignment, and image background removal on the acquired two-dimensional face image.

7. The method of claim 1 , wherein the three-dimensional reconstruction on the acquired two-dimensional face image is based on a TensorFlow-based colored 3D reconstruction algorithm, an Im2Avatar algorithm, a 3-SWEEP algorithm, or a 3D-GAN algorithm.

8. A face verification system, comprising a processor and a non-transitory computer-readable storage medium storing instructions executable by the processor to cause the system to perform operations comprising:

acquiring a two-dimensional face image and a three-dimensional face image for face verification of a target user;

performing face recognition on the acquired two-dimensional face image with respect to the target user to obtain a result of the face recognition;

in response to the result of the face recognition indicating that the face recognition is successful, performing three-dimensional reconstruction on the acquired two-dimensional face image to obtain a reconstructed three-dimensional face image;

performing liveness detection on the acquired three-dimensional face image to determine whether a face in the acquired three-dimensional face image corresponds to a live person;

in response to determining that the face in the acquired three-dimensional face image corresponds to a live person, comparing the reconstructed three-dimensional face image with the acquired three-dimensional face image to determine a similarity value; and

in response to the determined similarity value satisfying a condition, determining that the two-dimensional face image corresponds to the target user.

9. The system of claim 8 , wherein the comparing the reconstructed three-dimensional face image with the acquired three-dimensional face image to determine a similarity value comprises:

comparing three-dimensional structure data of the reconstructed three-dimensional face image with three-dimensional structure data of the acquired three-dimensional face image.

10. The system of claim 8 , wherein the condition comprises:

that the determined similarity value is not less than a threshold similarity value.

11. The system of claim 8 , wherein the performing face recognition on the acquired two-dimensional face image with respect to the target user to obtain a result of the face recognition comprises:

performing face recognition on the two-dimensional face image to obtain a face recognition value;

determining whether the face recognition value is not less than a threshold value;

obtaining a result indicating that the face recognition is successful if the face recognition value is not less than the threshold value; and

obtaining a result indicating that the face recognition is unsuccessful if the face recognition value is less than the threshold value.

12. The system of claim 8 , wherein the performing liveness detection on the acquired three-dimensional face image to determine whether a face in the acquired three-dimensional face image corresponds to a live person comprises:

performing liveness detection on the acquired three-dimensional face image to obtain a liveness detection value;

determining whether the liveness detection value is less than a threshold liveness value;

if the liveness detection value is not less than the threshold liveness value, determining that the face in the acquired three-dimensional face image corresponds to a live person; and

if the liveness detection value is less than the threshold liveness value, determining that the face in the acquired three-dimensional face image does not correspond to a live person.

13. The system of claim 8 , wherein the operations further comprise, before performing face recognition on the acquired two-dimensional face image:

subtracting an average brightness value from the acquired two-dimensional face image; or

performing face detection, face alignment, and image background removal on the acquired two-dimensional face image.

14. The system of claim 8 , wherein the three-dimensional reconstruction on the acquired two-dimensional face image is based on a TensorFlow-based colored 3D reconstruction algorithm, an Im2Avatar algorithm, a 3-SWEEP algorithm, or a 3D-GAN algorithm.

15. A non-transitory computer-readable storage medium for face verification, configured with instructions executable by one or more processors to cause the one or more processors to perform operations comprising:

acquiring a two-dimensional face image and a three-dimensional face image for face verification of a target user;

performing face recognition on the acquired two-dimensional face image with respect to the target user to obtain a result of the face recognition;

in response to the result of the face recognition indicating that the face recognition is successful, performing three-dimensional reconstruction on the acquired two-dimensional face image to obtain a reconstructed three-dimensional face image;

performing liveness detection on the acquired three-dimensional face image to determine whether a face in the acquired three-dimensional face image corresponds to a live person;

in response to determining that the face in the acquired three-dimensional face image corresponds to a live person, comparing the reconstructed three-dimensional face image with the acquired three-dimensional face image to determine a similarity value; and

in response to the determined similarity value satisfying a condition, determining that the two-dimensional face image corresponds to the target user.

16. The non-transitory computer-readable storage medium of claim 15 , wherein the comparing the reconstructed three-dimensional face image with the acquired three-dimensional face image to determine a similarity value comprises:

comparing three-dimensional structure data of the reconstructed three-dimensional face image with three-dimensional structure data of the acquired three-dimensional face image.

17. The non-transitory computer-readable storage medium of claim 15 , wherein the condition comprises:

that the determined similarity value is not less than a threshold similarity value.

18. The non-transitory computer-readable storage medium of claim 15 , wherein the performing face recognition on the acquired two-dimensional face image with respect to the target user to obtain a result of the face recognition comprises:

performing face recognition on the two-dimensional face image to obtain a face recognition value;

determining whether the face recognition value is not less than a threshold value;

obtaining a result indicating that the face recognition is successful if the face recognition value is not less than the threshold value; and

obtaining a result indicating that the face recognition is unsuccessful if the face recognition value is less than the threshold value.

19. The non-transitory computer-readable storage medium of claim 15 , wherein the performing liveness detection on the acquired three-dimensional face image to determine whether a face in the acquired three-dimensional face image corresponds to a live person comprises:

performing liveness detection on the acquired three-dimensional face image to obtain a liveness detection value;

determining whether the liveness detection value is less than a threshold liveness value;

if the liveness detection value is not less than the threshold liveness value, determining that the face in the acquired three-dimensional face image corresponds to a live person; and

if the liveness detection value is less than the threshold liveness value, determining that the face in the acquired three-dimensional face image does not correspond to a live person.

20. The non-transitory computer-readable storage medium of claim 15 , wherein the operations further comprise, before performing face recognition on the acquired two-dimensional face image:

subtracting an average brightness value from the acquired two-dimensional face image; or

performing face detection, face alignment, and image background removal on the acquired two-dimensional face image.