Granted Patent B1

US 12,450,945 · App. 18/915,483 · Granted Oct 21, 2025

Inventors: Chao Ma (Xi'an, CN); Pengtao Zhang (Xi'an, CN); Chao Wei (Xi'an, CN)

Assignee: Samsung Electronics Co., Ltd.

G06V40/172G06V10/751G06V40/45

Abstract

Claims (67)

1. A fingerprint image generating device, comprising:

processing circuitry, the processing circuitry configured to

obtain M sample fingerprint images;

generate N approximate images according to N target similarities based on each of the M sample fingerprint images, each of the N target similarities being a similarity value between a respective one of the N approximate images and a respective one of the M sample fingerprint images on which the respective one of the N approximate images is based, M and N being integers greater than 1;

analyze a difference between each one of the M sample fingerprint images and the respective one of the N approximate images, to obtain (M×N) difference parameters, each one of the (M×N) difference parameters quantifying a difference between each one of the respective M sample fingerprint images and the respective N approximate images, each one of the (M×N) difference parameters being associated with the similarity value between the respective one of the M sample fingerprint images and the respective one of the N approximate images, such that each of the N target similarities is associated with M difference parameters;

statistically calculate the M difference parameters respectively associated with P similarities among the N target similarities respectively to obtain difference statistical parameters respectively corresponding to the P similarities, P being an integer greater than or equal to 1;

obtain a reference fingerprint image; and

generate simulated fingerprint images respectively corresponding to the P similarities based on the reference fingerprint image and the difference statistical parameters respectively corresponding to the P similarities.

2. The fingerprint image generating device of claim 1 , wherein the processing circuitry is configured to:

obtain the N approximate images corresponding to the M sample fingerprint images by performing iterative calculations using an evolutionary algorithm, by respectively taking each of the M sample fingerprint images as a reference, wherein each of the N approximate images corresponds to one iteration; and

wherein each one of the N target similarities is represented by a number of iterations of the respective one of the N approximate images corresponding to an associated one of the M difference parameters.

3. The fingerprint image generating device of claim 2 , wherein, in the evolutionary algorithm, a structural similarity between a plurality of candidate images generated in each iteration and the reference is a fitness value of a corresponding candidate image of the plurality of candidate images, and the respective one of the N approximate images in a corresponding iteration is selected from the plurality of candidate images according to the fitness value.

4. The fingerprint image generating device of claim 2 , wherein the evolutionary algorithm comprises a genetic algorithm.

5. The fingerprint image generating device of claim 1 , wherein the processing circuitry is configured to:

respectively take each of the M sample fingerprint images and the respective one of the N approximate images as one image pair, to form (M×N) image pairs;

calculate differences in pixel values for each one of the (M×N) image pairs at corresponding pixel points as difference values; and

with respect to each one of the (M×N) image pairs, analyze a distribution rule of the difference values at each pixel point, and obtain a parameter describing the distribution rule as a (M×N) difference parameter of the respective one of the (M×N) image pairs to obtain the (M×N) difference parameters.

6. The fingerprint image generating device of claim 5 , wherein the processing circuitry is configured to:

with respect to each one of the (M×N) image pairs, perform frequency statistical analysis on the difference values at each pixel point, and obtain probability distribution parameters as the (M×N) difference parameter of the respective one of the (M×N) image pairs by a fitting operation to obtain the (M×N) difference parameters.

7. The fingerprint image generating device of claim 6 , wherein the probability distribution parameters comprise normal distribution parameters.

8. The fingerprint image generating device of claim 1 , wherein

the processing circuitry is configured to:

respectively statistically calculate the M difference parameters respectively associated with the N target similarities to obtain the difference statistical parameters of the N target similarities;

obtain a plurality of verified fingerprint images;

generate simulated fingerprint images respectively corresponding to the N target similarities according to the verified fingerprint images and the difference statistical parameters respectively corresponding to the N target similarities, with respect to each of the plurality of verified fingerprint images;

determine liveness scores of the simulated fingerprint images respectively corresponding to the N target similarities; and

select the P similarities from the N target similarities according to the liveness scores and a screening condition.

9. The fingerprint image generating device of claim 8 , wherein P is greater than 1, and the P similarities are a continuous interval in a similarity interval composed of the N target similarities; and

the screening condition includes, with respect to at least some verified fingerprint images, the liveness scores of the simulated fingerprint images corresponding to the P similarities continuously distributed over a liveness score interval.

10. A computing device, comprising:

at least one processor; and

at least one memory configured to store a computer program, the computer program, when executed by the at least one processor, is configured to

obtain M sample fingerprint images,

generate N approximate images according to N target similarities based on each of the M sample fingerprint images, each of the N target similarities being a similarity value between a respective one of the N approximate images and a respective one of the M sample fingerprint images on which the respective one of the N approximate images is based, M and N being integers greater than 1,

analyze a difference between each one of the M sample fingerprint images and the respective one of the N approximate images, to obtain (M×N) difference parameters, each one of the (M×N) difference parameters quantifying a difference between each one of the respective M sample fingerprint image and the respective N approximate images, each one of the (M×N) difference parameters being associated with the similarity value between the respective one of the M sample fingerprint images and the respective one of the N approximate images, such that each of the N target similarities is associated with M difference parameters,

statistically calculate the M difference parameters respectively associated with P similarities among the N target similarities respectively, to obtain difference statistical parameters respectively corresponding to the P similarities, P being an integer greater than or equal to 1,

obtain a reference fingerprint image, and

generate simulated fingerprint images respectively corresponding to the P similarities based on the reference fingerprint image and the difference statistical parameters respectively corresponding to the P similarities.

11. A fingerprint image generating method, comprising:

obtaining M sample fingerprint images;

generating N approximate images according to N target similarities based on each of the M sample fingerprint images, each of the N target similarities being a similarity value between a respective one of the N approximate images and a respective one of the M sample fingerprint images on which the respective one of the N approximate images is based, M and N being integers greater than 1;

analyzing a difference between each one of the M sample fingerprint images and the respective one of the N approximate images, to obtain (M×N) difference parameters, each one of the (M×N) difference parameters quantifying a difference between each one of the respective M sample fingerprint images and the respective N approximate images, each one of the (M×N) difference parameters being associated with the similarity value between the respective one of the M sample fingerprint images and the respective one of the N approximate images, such that each of the N target similarities is associated with M difference parameters;

statistically calculating the M difference parameters respectively associated with P similarities among the N target similarities respectively to obtain difference statistical parameters respectively corresponding to the P similarities, P being an integer greater than or equal to 1;

obtaining a reference fingerprint image; and

generating simulated fingerprint images respectively corresponding to the P similarities based on the reference fingerprint image and the difference statistical parameters respectively corresponding to the P similarities.

12. The fingerprint image generating method of claim 11 , wherein the generating of the N approximate images according to the N target similarities based on each of the M sample fingerprint images comprises:

obtaining the N approximate images corresponding to the M sample fingerprint images by performing iterative calculations using an evolutionary algorithm, by respectively taking each of the M sample fingerprint images as a reference, wherein each of the N approximate images corresponds to one iteration; and

wherein each one of the N target similarities is represented by a number of iterations of the respective one of the N approximate images corresponding to an associated one of the M difference parameters.

13. The fingerprint image generating method of claim 12 , wherein, in the evolutionary algorithm, a structural similarity between a plurality of candidate images generated in each iteration and the reference is a fitness value of a corresponding candidate image of the plurality of candidate images, and the respective one of the N approximate images in a corresponding iteration is selected from the plurality of candidate images according to the fitness value.

14. The fingerprint image generating method of claim 12 , wherein the evolutionary algorithm comprises a genetic algorithm.

15. The fingerprint image generating method of claim 11 , wherein the analyzing the difference between each of the M sample fingerprint images and the respective one of the N approximate images, to obtain the (M×N) difference parameters, comprises:

respectively taking each of the M sample fingerprint images and the respective one of the N approximate images as one image pair, to form (M×N) image pairs;

calculating differences in pixel values for each one of the (M×N) image pairs at corresponding pixel points as difference values; and

with respect to each one of the (M×N) image pairs, analyzing a distribution rule of the difference values at each pixel point, and obtaining a parameter describing the distribution rule as a (M×N) difference parameter of the respective one of the (M×N) image pairs to obtain the (M×N) difference parameters.

16. The fingerprint image generating method of claim 15 , wherein with respect to each one of the (M×N) image pairs, the analyzing the distribution rule of the difference values on each pixel point, and the obtaining the parameter describing the distribution rule as the (M×N) difference parameter, comprise:

with respect to each one of the (M×N) image pairs, performing frequency statistical analysis on the difference values at each pixel point, and obtaining probability distribution parameters as the (M×N) difference parameter of the respective one of the (M×N) image pairs by a fitting operation to obtain the (M×N) difference parameters.

17. The fingerprint image generating method of claim 16 , wherein the probability distribution parameters comprise normal distribution parameters.

18. The fingerprint image generating method of claim 11 , further comprising:

respectively statistically calculating the M difference parameters respectively associated with the N target similarities to obtain the difference statistical parameters of the N target similarities;

obtaining a plurality of verified fingerprint images;

with respect to each of the plurality of verified fingerprint images, generating simulated fingerprint images respectively corresponding to the N target similarities according to the verified fingerprint images and the difference statistical parameters respectively corresponding to the N target similarities;

determining liveness scores of the simulated fingerprint images respectively corresponding to the N target similarities; and

selecting the P similarities from the N target similarities according to the liveness scores and a screening condition.

19. The fingerprint image generating method of claim 18 , wherein P is greater than 1, and the P similarities are a continuous interval in a similarity interval composed of the N target similarities; and

the screening condition includes, with respect to at least some verified fingerprint images, the liveness scores of the simulated fingerprint images corresponding to the P similarities continuously distributed over a liveness score interval.

20. A non-transitory computer readable storage medium storing a computer program, wherein the computer program, when executed by a processor, is configured to implement the fingerprint image generating method of claim 11 .

21. A computer program product comprising instructions, wherein the instructions, when executed by a processor of a computer apparatus, implement the fingerprint image generating method of claim 11 .

Assignments (1)

Priority Claims (1)

References Cited (11)