An anti-counterfeiting method based on OCT internal and external fingerprint matching

Abstract

An anti-counterfeiting method based on internal and external fingerprint matching based on optical coherence tomography (OCT) comprises: 1) using an OCT device to collect finger OCT data; 2) separately obtaining an external fingerprint image and an internal fingerprint image, including automatic flattening of the curvature of the finger cross-section data, positioning of the internal and external fingerprint extraction areas, separately obtaining the internal and external fingerprints, linearly stretching the images, and enhancing the images, ultimately obtaining an external fingerprint image and an internal fingerprint image; 3) separately determining whether the external fingerprint image and the internal fingerprint image exist, i.e., using a fingerprint minutiae extractor to automatically check the minutiae of the external fingerprint image and the internal fingerprint image. If the external fingerprint image or the internal fingerprint image does not exist (the minutiae are 0), it is determined to be a fake forgery attack, otherwise it proceeds to the next step; 4) using a fingerprint matcher to perform matching. If the score is above a threshold, it is considered to be a real finger; otherwise, it is determined to be a fake forgery attack. The present invention provides an anti-counterfeiting method based on internal and external fingerprint matching based on OCT, which has high security, high anti-counterfeiting capability, and strong adaptability.

Description

Technical Field

[0001] The present invention relates to the field of biometric identity recognition technology, and in particular to a method for obtaining and matching inner and outer fingerprints of a finger based on OCT data to achieve anti-counterfeiting. Background Art

[0002] With the continuous development of information technology, biometric identification technology has begun to be applied in various industries. Widely used characteristics include but are not limited to DNA, fingerprints, retina, iris, facial features, earprints, lip prints, signatures, voiceprints, etc. Because biometric characteristics are unique, this characteristic makes biometric identification technology an important part of applications that require identity verification.

[0003] This is followed by the problem of forgery or impersonation in identity recognition technology. How to further strengthen the security of social informatization and networking has become a top priority that needs to be urgently resolved.

[0004] In recent years, fingerprint collection systems based on optical coherence tomography (OCT) have begun to emerge. OCT is a new, non-invasive, and non-destructive tomography technology with promising applications in the detection and imaging of living tissue. OCT can acquire three-dimensional internal information about a finger and extract both external and internal fingerprints from this three-dimensional information. Summary of the Invention

[0005] In order to overcome the shortcomings of existing identity authentication methods, such as poor security and susceptibility to attacks by forged samples, the present invention provides an anti-counterfeiting method based on OCT internal and external fingerprint matching, which is applicable to an OCT fingerprint system.

[0006] The technical solution adopted by the present invention to solve its technical problem is:

[0007] An anti-counterfeiting method based on OCT internal and external fingerprint matching includes the following steps:

[0008] 1) Use OCT equipment to collect finger OCT data;

[0009] 2) Obtaining external fingerprint images and internal fingerprint images respectively;

[0010] 3) Determine whether there is an external fingerprint image and an internal fingerprint image respectively. If the external fingerprint image or the internal fingerprint image does not exist, it is determined to be a prosthetic forgery attack, otherwise proceed to the next step;

[0011] 4) Use the fingerprint matcher to match. If the score is higher than the threshold, it is considered to be a real finger. Otherwise, it is judged as a prosthetic forgery attack.

[0012] Furthermore, in step 2), the fingerprint is generated by using the designated area pixel accumulation method. Each OCT finger cross-section image contains internal information of the finger, and the internal structure of the finger includes the stratum corneum, active epidermis, dermis, sweat glands, etc.

[0013] The following fingerprint extraction steps are used:

[0014] a) Automatic curvature flattening of finger cross-section data. First, the maximum gradient change is detected in each depth direction of the image, and the maximum value position is extracted. All pixels below the maximum pixel in the depth direction are sequentially moved to the fifth pixel in the new image to generate the curvature flattened image.

[0015] b) Internal and external fingerprint extraction area positioning. Each single OCT scan image is called a B-scan image. The areas are layered and positioned based on the internal structure of the finger in the B-scan image. The stratum corneum area is the external fingerprint extraction area, and the viable epidermis and dermis area are the internal fingerprint extraction area.

[0016] The definition of the external fingerprint depth area is shown in formula (1) (2):

[0017] Z ExtStart =1 (1)

[0018] Z ExtEnd =30 (2)

[0019] The definition of the internal fingerprint depth area is shown in formula (3) (4):

[0020] Z IntStart =30 (3)

[0021] Z IntEnd =200 (4)

[0022] c) Obtain internal and external fingerprints respectively, and define each pixel of the obtained fingerprint.

[0023] The pixel intensity of the external fingerprint is defined as shown in formula (5)

[0024]

[0025] The pixel intensity of the internal fingerprint is defined as shown in formula (6)

[0026]

[0027] Where I is the pixel value, x represents the abscissa of the B-scan image, y represents the yth consecutive B-scan image in the OCT finger volume data, and z represents the ordinate of the B-scan image.

[0028] d) Image linear stretching, mapping the generated fingerprint image pixel values ​​to 256 gray levels.

[0029] e) Image enhancement. The image is enhanced using the Multi-Scale Retinex algorithm and cropped and linearly stretched. For the external fingerprint image, the pixel values ​​between 50 and 127 are stretched to 0 and 255, and all other pixels are cut off. For the internal fingerprint image, the pixel values ​​between 50 and 178 are stretched to 0 and 255, and all other pixels are cut off. The resulting external and internal fingerprint images are obtained.

[0030] Furthermore, in step 3), the minutiae of the external and internal fingerprint images are automatically checked using a fingerprint minutiae extractor. If the number of minutiae is greater than 0, the fingerprint image is considered to be present. If no minutiae are present, it is considered a forgery attack. This method uses Verifinger to detect the number of minutiae in the fingerprint image.

[0031] Furthermore, in step 4), the Verifinger fingerprint matcher is used to match the fingerprint image to obtain a score. According to the matching rules of the Verifinger fingerprint matcher, the threshold is set to 50. If the score is higher than the threshold, it is considered to be a real finger, otherwise it is judged as a prosthetic forgery attack.

[0032] The technical concept of the present invention is: biometric technology is a technology for automatic identification of human biological characteristics (physiological or behavioral characteristics). Fingerprints, as a unique biological modality that everyone is born with, are currently the most mature and widely used biometric characteristics for identification. But at the same time, fingerprints are also a disaster area for the proliferation of counterfeit samples. Considering that the internal information of the finger contains an internal fingerprint with the same texture structure as the external fingerprint. Using the number of detail points of the internal fingerprint and the external fingerprint to detect can avoid common fake finger and fingerprint film attacks. For sufficiently thin fingerprint films, the attacker's external fingerprint information will also be retained after OCT acquisition, thereby generating a false internal fingerprint, but the detection of the matching score can resist such a false internal fingerprint.

[0033] This method fully utilizes the effective information from both internal and external fingerprints to extract and enhance both external and internal fingerprints, achieving high anti-counterfeiting capabilities and overcoming the vulnerability of external fingerprints to counterfeiting attacks. Furthermore, this anti-counterfeiting method is applicable to any OCT fingerprint acquisition system and can be used as a preprocessing method for OCT automatic fingerprint recognition systems, demonstrating its strong adaptability.

[0034] The beneficial effects of the present invention are mainly manifested in: high security, strong anti-counterfeiting ability and strong adaptability. BRIEF DESCRIPTION OF THE DRAWINGS

[0035] Figure 1 It is a flow chart of the method of the present invention.

[0036] Figure 2 This is a B-scan image of the three-dimensional internal information of the finger obtained by the OCT of the present invention.

[0037] Figure 3 This is a B-scan image of the finger cross-section data of the present invention with the curvature automatically flattened.

[0038] Figure 4 This is the B-scan image of the finger cross-section data curvature automatically flattened in the present invention, where the green line is Z ExtStart , the yellow line is Z ExtEnd .

[0039] Figure 5 This is the B-scan image of the finger cross-section data curvature automatically flattened in the present invention, where the blue line is Z IntStart , the red line is Z IntEnd .

[0040] Figure 6 It is the external fingerprint image mapped to 256 gray levels after linear stretching of the present invention.

[0041] Figure 7 It is the internal fingerprint image of the present invention mapped to 256 gray levels after linear stretching.

[0042] Figure 8 It is the external fingerprint image after image enhancement of the present invention.

[0043] Figure 9 It is the internal fingerprint image after image enhancement of the present invention.

[0044] Figure 10 This is the minutiae image obtained by extracting minutiae from an external fingerprint image using the Verifinger minutiae extractor of the present invention.

[0045] Figure 11 This is the minutiae image obtained by extracting minutiae from an internal fingerprint image using the Verifinger minutiae extractor of the present invention.

[0046] Figure 12 is the matching score obtained by using the Verifinger fingerprint matcher of the present invention. Specific implementation plan

[0047] The following is a further detailed description with reference to the accompanying drawings:

[0048] Reference Figure 1 , an anti-counterfeiting method based on OCT internal and external fingerprint matching, the process is as follows:

[0049] 1) Use OCT equipment to collect finger OCT data;

[0050] 2) Obtaining external fingerprint images and internal fingerprint images respectively;

[0051] 3) Determine whether there is an external fingerprint image and an internal fingerprint image respectively. If the external fingerprint image or the internal fingerprint image does not exist, it is determined to be a prosthetic forgery attack, otherwise proceed to the next step;

[0052] 4) Use the fingerprint matcher to match. If the score is higher than the threshold, it is considered to be a real finger. Otherwise, it is judged as a prosthetic forgery attack.

[0053] In step 2), the fingerprint is generated by using the designated area pixel accumulation method. Each OCT finger cross-section image contains the internal information of the finger, such as Figure 2 As shown, the internal structure of the finger includes the stratum corneum, active epidermis, dermis, sweat glands, etc.

[0054] The following fingerprint extraction steps are used:

[0055] a) Automatic flattening of the curvature of the finger cross-section data. First, perform a gradient maximum detection on each depth direction of the image and extract the maximum position. Then, move all pixels below the maximum pixel position in the depth direction to the fifth pixel of the new image in order to generate a curvature flattened image, as shown in the following example: Figure 3 shown.

[0056] b) Internal and external fingerprint extraction area positioning. Each single OCT scan image is called a B-scan image. The areas are layered and positioned based on the internal structure of the finger in the B-scan image. The stratum corneum area is the external fingerprint extraction area, and the viable epidermis and dermis area are the internal fingerprint extraction area.

[0057] The definition of the external fingerprint depth area is shown in formula (1) (2):

[0058] Z ExtStart =1 (1)

[0059] Z ExtEnd =30 (2)

[0060] Formula (1) represents Figure 4 The green line in formula (2) represents Figure 4 The yellow line in the

[0061] The definition of the internal fingerprint depth area is shown in formula (3) (4):

[0062] Z IntStart =30 (3)

[0063] Z IntEnd =200 (4)

[0064] Formula (3) represents Figure 5 The blue line in formula (4) represents Figure 5 The red line in.

[0065] c) Obtain internal and external fingerprints respectively, and define each pixel of the obtained fingerprint.

[0066] The pixel intensity of the external fingerprint is defined as shown in formula (5)

[0067]

[0068] The pixel intensity of the internal fingerprint is defined as shown in formula (6)

[0069]

[0070] Where I is the pixel value, x represents the abscissa of the B-scan image, y represents the yth consecutive B-scan image in the OCT finger volume data, and z represents the ordinate of the B-scan image.

[0071] d) Image linear stretching, mapping the generated fingerprint image pixel values ​​to 256 gray levels, such as Figure 6 As shown in , it is the external fingerprint after linear stretching. Figure 7 Shown is the inner fingerprint after linear stretching.

[0072] e) Image enhancement. The image is enhanced using the Multi-Scale Retinex algorithm and cropped and linearly stretched. The pixel values ​​of the external fingerprint image are stretched from 50-127 to 0-255, and the other pixels are cut off. The pixel values ​​of the internal fingerprint image are stretched from 50-178 to 0-255, and the other pixels are cut off. The final result is as follows Figure 8 The external fingerprint image shown and Figure 9 Internal fingerprint image shown.

[0073] In step 3), the minutiae of the external fingerprint image and the internal fingerprint image are automatically checked according to the fingerprint minutiae extractor. If the number of minutiae is greater than 0, it is determined that the fingerprint image exists. If no minutiae exist, it is determined to be a fake forgery attack. This method uses Verifinger to detect the number of minutiae of the fingerprint image, such as Figure 10 and Figure 11 shown.

[0074] In step 4), the Verifinger fingerprint matcher is used to match the fingerprint image to obtain a score, such as Figure 12As shown in the figure, the internal and external fingerprint matching scores of normal fingers will be higher than 50, so according to the matching rules of the Verifinger fingerprint matcher, the threshold is set to 50. If the score is higher than the threshold, it is considered to be a real finger, otherwise it is judged as a prosthetic forgery attack.

[0075] The above-mentioned embodiments are preferred embodiments of the present invention, but the embodiments of the present invention are not limited to the described embodiments. Any other modifications, substitutions, combinations, and tailorings that do not deviate from the spirit and principles of the present invention should be equivalent replacement methods and are included in the scope of protection of the present invention.

Claims

1. A method for anti-counterfeiting based on OCT internal and external fingerprint matching, comprising the following steps: 1) Use OCT equipment to collect finger OCT data; 2) Obtaining external fingerprint images and internal fingerprint images respectively; 3) Determine whether there is an external fingerprint image and an internal fingerprint image respectively. If the external fingerprint image or the internal fingerprint image does not exist, it is determined to be a prosthetic forgery attack, otherwise proceed to the next step; 4) Use the fingerprint matcher to match. If the score is higher than the threshold, it is considered to be a real finger. Otherwise, it is judged as a fake attack; In step 2), the fingerprint is generated by using a designated area pixel accumulation method; each pair of OCT finger cross-sectional images contains internal information of the finger, and the internal structure of the finger includes the stratum corneum, the active epidermis, the dermis, and the sweat glands; The following fingerprint extraction steps are used: a) Automatically flatten the curvature of the finger cross-section data. First, perform a gradient maximum detection in each depth direction of the image and extract the maximum value. Then, move all pixels below the maximum pixel in the depth direction to the fifth pixel in the new image in order to generate a curvature-flattened image. b) Positioning of internal and external fingerprint extraction areas: Each single OCT scan image is called a B-scan image. The areas are layered and positioned based on the internal structure of the finger in the B-scan image. The stratum corneum area is the external fingerprint extraction area, and the viable epidermis and dermis area are the internal fingerprint extraction area. The definition of the external fingerprint depth area is shown in formula (1) (2): Z ExtStart ＝1 (1) Z ExtEnd ＝30 (2) The definition of the internal fingerprint depth area is shown in formula (3) (4): Z IntStart ＝30 (3) Z IntEnd ＝200 (4) c) Obtain internal and external fingerprints respectively and define each pixel of the acquired fingerprint; The pixel intensity of the external fingerprint is defined as shown in formula (5) The pixel intensity of the internal fingerprint is defined as shown in formula (6) Where I is the pixel value, x represents the abscissa of the B-scan image, y represents the yth consecutive B-scan image in the OCT finger volume data, and z represents the ordinate of the B-scan image; d) Image linear stretching, mapping the generated fingerprint image pixel values ​​to 256 gray levels; e) Image enhancement: The image is enhanced using the Multi-Scale Retinex algorithm and cropped and linearly stretched. The pixel values ​​of the external fingerprint image are stretched to 0-255, and the other pixels are cut off. The pixel values ​​of the internal fingerprint image are stretched to 0-255, and the other pixels are cut off. Finally, the external fingerprint image and the internal fingerprint image are obtained. In step 3), the minutiae of the external fingerprint image and the internal fingerprint image are automatically checked according to the fingerprint minutiae extractor. If the number of minutiae is greater than 0, it is determined that the fingerprint image exists; if no minutiae exist, it is determined to be a prosthetic forgery attack; Verifinger is used to detect the number of minutiae of the fingerprint image; In step 4), the Verifinger fingerprint matcher is used to match the fingerprint image to obtain a score. According to the matching rules of the Verifinger fingerprint matcher, the threshold is set to 50; if the score is higher than the threshold, it is considered to be a real finger, otherwise it is judged to be a prosthetic forgery attack.

Images