KR-20260067781-A - METHOD AND APPARATUS FOR FAKE FINGERPRINTS GENERATION, AND TRAINING METHOD OF ARTIFICIAL INTELLIGENCE FOR IDENTIFYING FAKE FINGERPRINTS

KR20260067781AKR 20260067781 AKR20260067781 AKR 20260067781AKR-20260067781-A

Abstract

A method for training a forged fingerprint detection model according to the first aspect of the present invention comprises the steps of: acquiring a physical forged fingerprint image generated physically with a predetermined finger as a target; acquiring a training fingerprint image generated to share unique identification information possessed by the actual fingerprint with the actual fingerprint; and training a predetermined forged fingerprint detection model using the physical forged fingerprint image and the training fingerprint image in a transfer learning manner or in an ensemble learning manner. In this case, the training fingerprint image is acquired by providing the unique identification information and generation constraints aimed at diversifying the generated training fingerprint image to a predetermined training fingerprint image generation model.

Inventors

김범준
이종만
신호철
박재현
강영묵

Assignees

주식회사 슈프리마

Dates

Publication Date: 20260513
Application Date: 20241106

Claims (20)

In a method for learning a forged fingerprint detection model performed by a learning device of a forged fingerprint detection model, A step of acquiring a physical forged fingerprint image generated physically using a specific finger as a target; A step of acquiring a training fingerprint image generated to share unique identification information possessed by an actual fingerprint with said actual fingerprint; and The method includes the step of training a predetermined forged fingerprint detection model using the physical forged fingerprint image and the training fingerprint image in a transfer learning manner or in an ensemble learning manner, wherein The above-mentioned learning fingerprint image is, The one obtained by providing generation constraints aimed at diversifying the generated training fingerprint images and the unique identification information to a predetermined training fingerprint image generation model Training method for a forged fingerprint detection model.
In paragraph 1, In the learning of the above-mentioned transfer learning method, The above physical forged fingerprint image is used for pre-training and the above training fingerprint image is used for fine-tuning Training method for a forged fingerprint detection model.
In paragraph 2, In the aforementioned prior learning, An image with masking applied to a portion of the physical forged fingerprint image is provided as a training input, and an image without masking is provided as a training answer. Training method for a forged fingerprint detection model.
In paragraph 2, The above forged fingerprint detection model includes a physical forged fingerprint detection model and a training forged fingerprint detection model, and In the above-mentioned ensemble learning method, the above-mentioned physical forged fingerprint detection model and the above-mentioned training forged fingerprint detection model are each trained, The above physical forgery fingerprint detection model is, The above-mentioned physical forged fingerprint image is provided as a training input, and whether the provided image is forged is provided as a training answer to perform training. The above-mentioned forged fingerprint detection model for training is, The above-mentioned training fingerprint image is provided as a training input, and whether the provided image is forged is provided as a training answer to perform training. Training method for a forged fingerprint detection model.
In paragraph 1, The above unique identification information is, at least one of first identification information based on the start and end points and branching points of a ridge within a fingerprint of a finger, second identification information based on the direction of the ridge, and third identification information based on the distribution density of the ridge. Training method for a forged fingerprint detection model.
In paragraph 1, The above generation constraint is, Information regarding a sensor for identifying a fingerprint, information regarding the surrounding environment in which the fingerprint is identified, and labeling information regarding the fingerprint Training method for a forged fingerprint detection model.
In paragraph 6, The above labeling information is, Information regarding the type of physical means used to generate the above-mentioned physical forged fingerprint image Training method for a forged fingerprint detection model.
In paragraph 1, The above-mentioned fingerprint image generation model for training is, The one trained to generate different training fingerprint images according to conditions included in the generation constraints while the above-mentioned unique identification information is frozen. Training method for a forged fingerprint detection model.
In paragraph 1, The above-mentioned fingerprint image generation model for training is, Acquire random noise and generate different training fingerprint images according to the random noise even if the generation constraints are the same. Training method for a forged fingerprint detection model.
Memory capable of storing computer-executable instructions; and By executing the above command, a physical forged fingerprint image generated physically targeting a specific finger is obtained, and Acquire a training fingerprint image generated to share the unique identification information of the actual fingerprint with the actual fingerprint, and A processor comprising, using the physical forged fingerprint image and the training fingerprint image, training a predetermined forged fingerprint detection model using a transfer learning method or an ensemble learning method, The above-mentioned learning fingerprint image is, The one obtained by providing generation constraints aimed at diversifying the generated training fingerprint images and the unique identification information to a predetermined training fingerprint image generation model Learning device for a forged fingerprint detection model.
In Paragraph 10, In the learning of the above-mentioned transfer learning method, The above physical forged fingerprint image is used for pre-training and the above training fingerprint image is used for fine-tuning Learning device for a forged fingerprint detection model.
In Paragraph 11, In the aforementioned prior learning, An image with masking applied to a portion of the physical forged fingerprint image is provided as a training input, and an image without masking is provided as a training answer. Learning device for a forged fingerprint detection model.
In Paragraph 11, The above forged fingerprint detection model includes a physical forged fingerprint detection model and a training forged fingerprint detection model, and In the above-mentioned ensemble learning method, the above-mentioned physical forged fingerprint detection model and the above-mentioned training forged fingerprint detection model are each trained, The above physical forgery fingerprint detection model is, The above-mentioned physical forged fingerprint image is provided as a training input, and whether the provided image is forged is provided as a training answer to perform training. The above-mentioned forged fingerprint detection model for training is, The above-mentioned training fingerprint image is provided as a training input, and whether the provided image is forged is provided as a training answer to perform training. Learning device for a forged fingerprint detection model.
In Paragraph 10, The above unique identification information is, at least one of first identification information based on the start and end points and branching points of a ridge within a fingerprint of a finger, second identification information based on the direction of the ridge, and third identification information based on the distribution density of the ridge. Learning device for a forged fingerprint detection model.
In Paragraph 10, The above generation constraint is, Information regarding a sensor for identifying a fingerprint, information regarding the surrounding environment in which the fingerprint is identified, and labeling information regarding the fingerprint Learning device for a forged fingerprint detection model.
In paragraph 15, The above labeling information is, Information regarding the type of physical means used to generate the above-mentioned physical forged fingerprint image Learning device for a forged fingerprint detection model.
In Paragraph 10, The above-mentioned fingerprint image generation model for training is, The one trained to generate different training fingerprint images according to conditions included in the generation constraints while the above-mentioned unique identification information is frozen. Learning device for a forged fingerprint detection model.
In Paragraph 10, The above-mentioned fingerprint image generation model for training is, Acquire random noise and generate different training fingerprint images according to the random noise even if the generation constraints are the same. Learning device for a forged fingerprint detection model.
A computer-readable recording medium storing computer-executable instructions, wherein the computer-executable instructions, when executed by a processor, A step of acquiring a physical forged fingerprint image generated physically using a specific finger as a target; A step of acquiring a training fingerprint image generated to share unique identification information possessed by an actual fingerprint with said actual fingerprint; and The method includes the step of training a predetermined forged fingerprint detection model using the physical forged fingerprint image and the training fingerprint image in a transfer learning manner or in an ensemble learning manner, wherein The above-mentioned learning fingerprint image is, A method in which the processor performs the method obtained by providing generation constraints aimed at diversifying the generated training fingerprint images and the unique identification information to a predetermined training fingerprint image generation model. Computer-readable recording medium.
As a computer program stored on a computer-readable recording medium, When the above computer program is executed by a processor, A step of acquiring a physical forged fingerprint image generated physically using a specific finger as a target; A step of acquiring a training fingerprint image generated to share unique identification information possessed by an actual fingerprint with said actual fingerprint; and The method includes the step of training a predetermined forged fingerprint detection model using the physical forged fingerprint image and the training fingerprint image in a transfer learning manner or in an ensemble learning manner, wherein The above-mentioned learning fingerprint image is, Instructions for the processor to perform a method obtained by providing generation constraints aimed at diversifying the generated training fingerprint images and the unique identification information to a predetermined training fingerprint image generation model. A computer program stored on a computer-readable recording medium.

Description

Method and apparatus for generating fake fingerprints, and method of training an artificial intelligence model for identifying fake fingerprints The present invention relates to a method and apparatus for generating a forged fingerprint, and a method for training an artificial intelligence model for detecting a forged fingerprint. Fingerprint recognition is one of the most widely used biometric technologies in modern times, rapidly replacing passwords consisting of characters or patterns in everyday life. While many people perceive fingerprint technology as being more difficult to leak or duplicate than characters or patterns, fingerprints are not immune to duplication; therefore, if a fingerprint is exposed to others and duplicated, authentication could still be possible using a forged, duplicated fingerprint. Meanwhile, fingerprints can pose a security vulnerability as they are easily exposed to the outside world, such as being left on every object touched. Accordingly, the development of technologies to distinguish between real and forged fingerprints (for example, the development of forged fingerprint detection technology using artificial intelligence) is actively underway. However, there are limitations in improving the performance of forged fingerprint detection models due to insufficient training data to train conventional AI algorithms for identifying forged fingerprints. FIG. 1 is a block diagram exemplarily illustrating a forged fingerprint generating device according to one embodiment. Figure 2 is a block diagram exemplifying the function of a forged fingerprint generation program. Figure 3 is an example diagram showing the relationship between the unique identification information extraction unit, the training fingerprint image acquisition unit, and the model training unit of a forged fingerprint generation program. FIG. 4 is a flowchart exemplarily showing the operation of the model learning unit of a forged fingerprint generation program according to one embodiment. FIG. 5 is a flowchart exemplarily showing the operation of a learning fingerprint image acquisition unit of a forged fingerprint generation program according to one embodiment. FIG. 6 is a flowchart exemplarily illustrating a method for generating a forged fingerprint according to one embodiment. FIG. 7 is an example diagram showing, in its entirety, a method for training a fingerprint image generation model according to one embodiment and a method for generating a fingerprint image for training by the model thus trained. FIG. 8 is a block diagram exemplarily illustrating a forged fingerprint detection device according to one embodiment. Figure 9 is a block diagram exemplifying the function of a forged fingerprint detection program. FIG. 10 is a flowchart exemplarily showing a learning method of a forged fingerprint detection model according to one embodiment. FIGS. 11 to 14 are exemplary diagrams showing a method of training a forged fingerprint detection model according to different embodiments. The advantages and features of the present invention and the methods for achieving them will become clear by referring to the embodiments described below in detail together with the accompanying drawings. However, the present invention is not limited to the embodiments disclosed below but can be implemented in various different forms. These embodiments are provided merely to ensure that the disclosure of the present invention is complete and to fully inform those skilled in the art of the scope of the invention, and the present invention is defined only by the scope of the claims. In describing the embodiments of the present invention, specific descriptions of known functions or configurations will be omitted if it is determined that such detailed descriptions could unnecessarily obscure the essence of the invention. Furthermore, the terms described below are defined in consideration of their functions in the embodiments of the present invention, and these definitions may vary depending on the intentions or practices of the user or operator. Therefore, such definitions should be based on the content throughout this specification. The terms used in this specification will be briefly explained, and the invention will be described in detail. The terms used in this specification have been selected to be as widely used as possible, taking into account the functions of the present invention; however, these terms may vary depending on the intent of those skilled in the art, case law, the emergence of new technologies, etc. Additionally, in specific cases, terms have been arbitrarily selected by the applicant, and in such cases, their meanings will be described in detail in the relevant description of the invention. Therefore, the terms used in this invention should be defined not merely by their names, but based on their meanings and the overall content of the invention. When a part of a specification is described as 'comprising' a certain compo