Fingerprint extraction apparatus and method

1. A fingerprint extraction apparatus comprising:

a fingerprint generation module configured to generate at least one first virtual fingerprint image comprising a fingerprint located vertically on a center of an image, and a plurality of second virtual fingerprint images different from the at least one first virtual fingerprint image, perform primary image processing on the at least one first virtual fingerprint image, perform primary image processing and secondary image processing on each of the plurality of second virtual fingerprint images, and combine the at least one first virtual fingerprint image on which the primary image processing is performed and the plurality of second virtual fingerprint images on which the at least one of primary image processing and secondary image processing is performed to generate a plurality of virtual overlapped fingerprint images;

a machine learning module configured to generate a learning model for extracting a target fingerprint by performing machine learning using the plurality of virtual overlapped fingerprint images as input data and the at least one first virtual fingerprint image as output data; and

a fingerprint extraction module configured to extract a fingerprint located vertically on a center of a real image by inputting the real image to the learning model,

wherein the primary image processing comprises image processing on a curve forming a fingerprint, and the secondary image processing comprises image processing on a location of the fingerprint in the image, and

wherein the primary image processing further comprises at least one of adding a curve of the fingerprint, removing a curve of the fingerprint, and changing a thickness of the fingerprint, and the secondary image processing comprises at least one of rotating the fingerprint and inverting the fingerprint.

2. The fingerprint extraction apparatus of claim 1 , wherein the fingerprint generation module is further configured to generate a plurality of virtual background images, and generate a plurality of virtual overlapped fingerprint images by combining the at least one first virtual fingerprint image on which the primary image processing is performed, the plurality of second virtual fingerprint images on which the at least one of primary image processing and secondary image processing is performed, and the plurality of virtual background images, and

wherein each of the plurality of virtual background images comprises at least one pattern other than the fingerprint.

3. The fingerprint extraction apparatus of claim 1 , wherein each of the plurality of second virtual fingerprint images comprises a fingerprint that is not located on the center of the image or is not located vertically on the image.

4. The fingerprint extraction apparatus of claim 1 , wherein a shape of the fingerprint included in each of the plurality of second virtual fingerprint images is different from a shape of the fingerprint included in the at least one first virtual fingerprint image.

5. The fingerprint extraction apparatus of claim 1 , wherein a shape of the fingerprint included in each of the plurality of second virtual fingerprint images is the same as a shape of the fingerprint included in the at least one first virtual fingerprint image.

6. The fingerprint extraction apparatus of claim 1 , wherein the fingerprint extraction module is further configured to determine a target fingerprint from the real image, pre-process the real image such that the target fingerprint is vertically located on the center of the real image, and input the pre-processed real image into the learning model to extract the target fingerprint.

7. A fingerprint extraction method comprising:

generating at least one first virtual fingerprint image comprising a fingerprint located vertically on a center of an image;

generating a plurality of second virtual fingerprint images different from the at least one first virtual fingerprint image;

performing primary image processing on the at least one first virtual fingerprint image;

performing primary image processing and secondary image processing on each of the plurality of second virtual fingerprint images;

generating a plurality of virtual overlapped fingerprint images by combining the at least one first virtual fingerprint image on which the primary image processing is performed and the plurality of second virtual fingerprint images on which the at least one of primary image processing and secondary image processing is performed;

generating a learning model for extracting a target fingerprint by performing machine learning using the plurality of virtual overlapped fingerprint images as input data and the at least one first virtual fingerprint image as output data; and

extracting a fingerprint located vertically on a center of a real image by inputting the real image to the learning model,

wherein the primary image processing comprises image processing on a curve forming a fingerprint, and the secondary image processing comprises image processing on a location of the fingerprint in the image, and

wherein the primary image processing further comprises at least one of adding a curve of the fingerprint, removing a curve of the fingerprint, and changing a thickness of the fingerprint, and the secondary image processing comprises at least one of rotating the fingerprint and inverting the fingerprint.

8. The fingerprint extraction method of claim 7 , further comprising generating a plurality of virtual background images,

wherein the generating of the plurality of virtual overlapped fingerprint images comprises generating the plurality of virtual overlapped fingerprint images by combining the at least one first virtual fingerprint image on which the primary image processing is performed, the plurality of second virtual fingerprint images on which the at least one of primary image processing and secondary image processing is performed, and the plurality of virtual background images, and

wherein each of the plurality of virtual background images comprises at least one pattern other than the fingerprint.

9. The fingerprint extraction method of claim 7 , wherein each of the plurality of second virtual fingerprint images comprises a fingerprint that is not located on the center of the image or is not located vertically on the image.

10. The fingerprint extraction method of claim 7 , wherein a shape of the fingerprint included in each of the plurality of second virtual fingerprint images is different from a shape of the fingerprint included in the at least one first virtual fingerprint image.

11. The fingerprint extraction method of claim 7 , wherein a shape of the fingerprint included in each of the plurality of second virtual fingerprint images is the same as a shape of the fingerprint included in the at least one first virtual fingerprint image.

12. The fingerprint extraction method of claim 7 , wherein the extracting of the fingerprint comprises:

determining a target fingerprint from the real image;

pre-processing the real image such that the target fingerprint is vertically located on the center of the real image; and

inputting the pre-processed real image into the learning model to extract the target fingerprint.