Within document face verification

1. A computer-implemented method for determining whether images of faces contained in a document correspond to each other, the method comprising:

acquiring image data pertaining to the document; and

performing facial detection on the image data to detect one or more facial representations existing within the document, wherein performing facial detection on the image data comprises using at least one face detection algorithm and wherein:

if two or more facial representations are detected:

selecting a first facial representation and a second facial representation; and

determining whether the image of a face of the first facial representation corresponds to the image of a face of the second facial representation; and

if fewer than two facial representation are detected, identifying the document as being invalid.

2. The computer-implemented method of claim 1 , wherein acquiring image data pertaining to the document comprises acquiring at least one image of at least a part of the document.

3. The computer-implemented method of claim 2 , wherein the at least one image is acquired using a visible-light camera, an ultraviolet camera, or an infrared camera, or by transferring from a memory.

4. The computer-implemented method of claim 2 , further comprising manipulating the at least one image to obtain image data pertaining to the document,

optionally comprising one or more of: applying one or more image transforms to the at least one image, cropping the at least one image, and forming a single document image from the at least one image.

5. The computer-implemented method of claim 4 , wherein performing facial detection on the image data comprises using a first face detection algorithm and a second face detection algorithm on the image data,

optionally wherein the first face detection algorithm is configured to detect the first facial representation and the second face detection algorithm is configured to detect the second facial representation.

6. The computer-implemented method of claim 1 , wherein each of the one or more facial representation has one or more of the following computer-readable characteristics:

the facial representation is formed of alphanumeric characters;

the facial representation is formed of pixels, optionally wherein the pixels are pixels are varying shades, greyscale, or colour;

the facial representation is a hologram;

the facial representation is visible under visible light;

the facial representation is visible under ultraviolet light only; and

the facial representation is visible under infrared light only.

7. The computer-implemented method of claim 1 , wherein selecting a first facial representation and a second facial representation is performed automatically based on one or more detected computer-readable characteristics of the detected facial representations.

8. The computer-implemented method of claim 1 , wherein determining whether the image of a face of the first facial representation corresponds to the image of a face of the second facial representation comprises:

extracting at least one feature from each of the first facial representation and the second facial representation;

combining the extracted features to produce a pair-feature representation; and

determining whether the image of a face of the first facial representation corresponds to the image of a face of the second facial representation using the pair-feature representation.

9. The computer-implemented method of claim 8 , further comprising determining at least one feature for extraction from each of the first facial representation and the second facial representation via a computer vision algorithm or a deep neural network.

10. The computer-implemented method of claim 8 , wherein combining the extracted features comprises one or more of:

concatenating the extracted features of the first facial representation with the extracted features of the second facial representation;

adding the extracted features of the first facial representation to the extracted features of the second facial representation;

subtracting the extracted features of the second facial representation from the extracted features of the first facial representation; and

subtracting the extracted features of the first facial representation from the extracted features of the second facial representation.

11. The computer-implemented method of claim 1 , further comprising prior to the determining:

applying at least one transform to at least one of the first facial representation and second facial representation such that the first facial representation and the second facial representation are in the same representation space.

12. The computer-implemented method of claim 11 , wherein a first transform is applied to the first facial representation to transform the first facial representation into the representation space of the second facial representation, or

wherein a second transform is applied to the second facial representation to transform the second facial representation into the representation space of the first facial representation, or

wherein a first transform is applied to the first facial representation and a second transform is applied to the second facial representation to transform both the first facial representation and the second facial representation into a representation space different from the representation space of the first facial representation and the representation space of the second facial representation.

13. A computer-implemented method for testing the authenticity of a document comprising the method of claim 1 , and further comprising:

determining whether the document is authentic based on whether the image of a face in the first facial representation corresponds to the image of a face in the second facial representation.

14. A computer comprising a processor configured to execute executable code stored in memory, wherein the executable code comprises instructions for performing a method comprising:

acquiring image data pertaining to the document; and

performing facial detection on the image data to detect one or more facial representations existing within the document, wherein performing facial detection on the image data comprises using at least one face detection algorithm and wherein:

if two or more facial representations are detected:

selecting a first facial representation and a second facial representation; and

determining whether the image of a face of the first facial representation corresponds to the image of a face of the second facial representation; and

if fewer than two facial representations are detected, identifying the document as being invalid.