Perspective distortion characteristic based facial image authentication method and storage and processing device thereof

A storage device storing a plurality of programs, wherein the programs are capable of being loaded by a processor to execute a perspective distortion characteristic based facial image authentication method comprising

• step S1:
• recognizing key points and a contour in a two-dimensional facial image;
• step S2:
• acquiring key points in a three-dimensional facial model based on the three-dimensional facial model corresponding to the two-dimensional facial image;
• step S3:
• calculating camera parameters based on a correspondence between the key points in the two-dimensional facial image and the key points in the three-dimensional facial model;
• step S4:
• optimizing the camera parameters obtained in step S3 based on the contour in the two-dimensional facial image;
• step S5:
• randomly sampling the key points in the two-dimensional facial image, and repeating steps S3 and S4 until a preset loop condition is satisfied, and obtaining a camera intrinsic parameter estimation point cloud according to the camera parameters acquired in step S4 in each loop;
• and
• step S6:
• calculating an inconsistency between the camera intrinsic parameter estimation point cloud and camera nominal intrinsic parameters, and determining an authenticity of the two-dimensional facial image based on the inconsistency;
• wherein the camera nominal intrinsic parameters are parameters of a camera capturing the two-dimensional facial image,
• wherein a method for calculating the camera parameters in step S3, comprises:
• step S31:
• calculating a camera projection matrix by using a gold standard method based on the key points in the two-dimensional facial image and the key points in the three-dimensional facial model;
• and
• step S32:
• based on the camera projection matrix calculated in step S31, solving for the camera parameters of 9 degrees of freedom by adding a constraint that a pixel unit is square;
• the camera parameters of 9 degrees of freedom comprise camera intrinsic parameters of 3 degrees of freedom and camera extrinsic parameters of 6 degrees of freedom.

A processing device comprising:

• a processor, capable of executing a plurality of programs;
• and
• a storage device, capable of storing the plurality of programs;
• wherein the programs are capable of being loaded by the processor to execute a perspective distortion characteristic based facial image authentication method comprising
• step S1:
• recognizing key points and a contour in a two-dimensional facial image;
• step S2:
• acquiring key points in a three-dimensional facial model based on the three-dimensional facial model corresponding to the two-dimensional facial image;
• step S3:
• calculating camera parameters based on a correspondence between the key points in the two-dimensional facial image and the key points in the three-dimensional facial model;
• step S4:
• optimizing the camera parameters obtained in step S3 based on the contour in the two-dimensional facial image;
• step S5:
• randomly sampling the key points in the two-dimensional facial image, and repeating steps S3 and S4 until a preset loop condition is satisfied, and obtaining a camera intrinsic parameter estimation point cloud according to the camera parameters acquired in step S4 in each loop;
• and
• step S6:
• calculating an inconsistency between the camera intrinsic parameter estimation point cloud and camera nominal intrinsic parameters, and determining an authenticity of the two-dimensional facial image based on the inconsistency;
• wherein the camera nominal intrinsic parameters are parameters of a camera capturing the two-dimensional facial image,
• wherein a method for calculating the camera parameters in step S3, comprises:
• step S31:
• calculating a camera projection matrix by using a gold standard method based on the key points in the two-dimensional facial image and the key points in the three-dimensional facial model;
• and
• step S32:
• based on the camera projection matrix calculated in step S31, solving for the camera parameters of 9 degrees of freedom by adding a constraint that a pixel unit is square;
• the camera parameters of 9 degrees of freedom comprise camera intrinsic parameters of 3 degrees of freedom and camera extrinsic parameters of 6 degrees of freedom.