Search

EP-3819818-B1 - ELECTRONIC DEVICE AND METHOD FOR OBTAINING FEATURES OF BIOMETRICS

EP3819818B1EP 3819818 B1EP3819818 B1EP 3819818B1EP-3819818-B1

Inventors

  • LIAO, CHIEN-FU

Dates

Publication Date
20260506
Application Date
20200218

Claims (13)

  1. An electronic device (100) for biometric verification of a user using biometric features derived from an image (300) of a finger of the user, comprising: an input circuit (103); and a processor (101) coupled to the input circuit (103), wherein the input circuit (103) is configured to obtain the image (300) of the finger of the user including a fingerprint and a finger vein at the same time, the processor (101) is configured to performing a pre-processing and a binarization processing on the image (300) to obtain a binarized image and performing a mathematical morphology operation on the binarized image to obtain a plurality of feature points in the image (300) , wherein the mathematical morphology operation is an image analysis method based on a mathematical set theory; divide the image (300) into a plurality of sub-images (SIMG), wherein the sub-images do not overlap with each other and have the same size and each of the sub-images (SIMG) comprises a plurality of first feature points among the plurality of feature points, obtain, for each sub-image of the plurality sub-images, a specific region (R1) and other regions in the sub-image (SIMG) via a particle swarm optimization algorithm, wherein the specific region (R1) comprises a plurality of second feature points of the first feature points and a distribution density of the second feature points in the specific region (R1) is higher than in any other regions of the sub-image, identify a plurality of third feature points from the second feature points by selecting, from the second feature points, the feature points having a number of occurrences greater than a threshold value as the third feature points for performing the biometric verification operation, comprising: obtaining a plurality of optimal sets including a plurality of fourth feature points in other finger images that are for performing the biometric verification operation, and selecting, from the second feature points, the feature points having the number of occurrences greater than the threshold value as the third feature points, wherein the number of occurrences of the feature points is associated with an appearing frequency of the feature points across the second feature points and the fourth feature points, and perform the biometric verification operation of the user of the electronic device based on the third feature points.
  2. The electronic device as claimed in claim 1, wherein after identifying the plurality of third feature points for performing the biometric operation, the processor (101) is configured to determine whether the number of the plurality of third feature points is greater than a second threshold value, when the number of the plurality of third feature points is greater than the second threshold value, to select a plurality of fifth feature points from the plurality of third feature points by performing a sequential forward selection and to perform the biometric verification operation of the user of the electronic device based on the fifth feature points, wherein the number of the plurality of fifth feature points is smaller than the number of the plurality of third feature points.
  3. The electronic device as claimed in any of claims 1 to 2, wherein the fingerprint is characterized as a first type biological feature, the finger vein is characterized as a second type biological feature, and the electronic device further comprises: a storage circuit (115) coupled to the processor (101) and comprising a plurality of secure isolated regions, wherein the processor (101) is configured to store among the plurality of third feature points the feature points belonging to the first type biological feature in a second storage region of the secure isolated regions corresponding to the second type biological feature and to store among the plurality of third feature points the feature points belonging to the second type biological feature in a first storage region of the secure isolated regions corresponding to the first type biological feature in the storage circuit (115), wherein the electronic device further comprises: a communication circuit coupled to the processor (101), wherein the processor (101) is configured to obtain a first public key, a first private key corresponding to the first public key, a second public key and a second private key corresponding to the second public key, to obtain a plurality of encrypted data received by the communication circuit comprising first encrypted data encrypted by the first public key and second encrypted data encrypted by the second public key, to decrypt the first encrypted data by the first private key to obtain a first to-be-verified feature point of the first type biological feature, to decrypt the second encrypted data by the second private key to obtain a second to-be-verified feature point of the second type biological feature, and to perform the biometric verification operation using the first to-be-verified feature point and the second to-be-verified feature point, wherein in performing the biometric verification operation using the first to-be-verified feature point and the second to-be-verified feature point the processor (101) is further configured to obtain first matching data with the highest similarity from the second storage region according to the first to-be-verified feature point, to obtain second matching data with the highest similarity from the first storage region according to the second to-be-verified feature point, and determine whether an index of the first matching data is the same as an index of the second matching data, wherein, in response to the index of the first matching data being the same as the index of the second matching data, the processor (101) is configured to determine that the plurality of encrypted data have passed verification, and in response to the index of the first matching data not being the same as the index of the second matching data, the processor (101) is configured to determine that the plurality of encrypted data have not passed the verification.
  4. The electronic device as claimed in claim 3, wherein before determining whether the index of the first matching data is the same as the index of the second matching data, the processor (101) is configured to determine whether the first matching data and the second matching data meet a preset normalization standard, , in response to determining that a similarity between the first matching data and the first to-be-verified feature point is greater than a threshold value, to determine that the first matching data meets the preset normalization standard, and in response to determining that a similarity between the second matching data and the second to-be-verified feature point is greater than the threshold value, to determine that the second matching data meets the preset normalization standard; , in response to at least one of the first matching data and the second matching data not meeting the preset normalization standard, to determine that the plurality of encrypted data have not passed the verification, and in response to both the first matching data and the second matching data meeting the preset normalization standard, to determine whether the index of the first matching data is the same as the index of the second matching data.
  5. The electronic device as claimed in claim 1, wherein the fingerprint is characterized as a first type biological feature, the finger vein is characterized as a second type biological feature, and the electronic device further comprises: a storage circuit (115) coupled to the processor (101), wherein the processor (101) is configured to cross-combine the feature points among the plurality of third feature points that belong to the first type biological feature with the feature points among the plurality of third feature points that belong to the second type biological feature to generate combined data, to perform a rail fence encryption algorithm on the combined data and obtain first encrypted data, and to store the first encrypted data in the storage circuit (115), to obtain a third to-be-verified feature point of the first type biological feature and a fourth to-be-verified feature point of the second type biological feature, to decrypt at least one encrypted data stored in the storage circuit (115) and obtains at least one decrypted data, to compare the decrypted data with the third to-be-verified feature point and the fourth to-be-verified feature point and obtain third matching data with the highest similarity in the first type biological feature and fourth matching data with the highest similarity in the second type biological feature, to determine whether the third matching data and the fourth matching data meet a preset normalization standard, , in response to determining that a similarity between the third matching data and the third to-be-verified feature point is greater than a threshold value, to determine that the third matching data meets the preset normalization standard, and in response to determining that a similarity between the fourth matching data and the fourth to-be-verified feature point is greater than the threshold value, to determine that the fourth matching data meets the preset normalization standard; , in response to at least one of the third matching data and the fourth matching data not meeting the preset normalization standard, to determine that the biometric verification operation is not passed, and in response to both of the first matching data and the second matching data meeting the preset normalization standard, to determine that the biometric verification operation is passed.
  6. The electronic device as claimed in claim 5, wherein before determining whether the third matching data and the fourth matching data meet the preset normalization standard, the processor (101) is configured to determine whether an index of the third matching data is the same as an index of the fourth matching data, , in response to the index of the third matching data being the same as the index of the fourth matching data, to determine whether the third matching data and the fourth matching data meet the preset normalization standard, and , in response to the index of the third matching data not being the same as the index of the fourth matching data, to determine that the biometric operation is not passed.
  7. The electronic device as claimed in any of the preceding claims, further comprising a motor circuit (109) coupled to the processor (101), wherein the motor circuit (109) is configured to receive an indication signal from the processor (101) based on the biometric verification operation of the user of the electronic device to determine whether to lock or unlock the electronic device (100).
  8. A method for biometric verification of a user using biometric features derived from an image (300) of a finger of the user, adapted to an electronic device comprising an input circuit (103) and a processor (101) coupled to the input circuit (103), the method comprising: obtaining, by the input circuit (103), the image of the finger of the user including a fingerprint and a finger vein at the same time; performing, by the processor, a pre-processing and a binarization processing on the image (300) to obtain a binarized image, and performing a mathematical morphology operation on the binarized image to obtain a plurality of feature points in the image (300), wherein the mathematical morphology operation is an image analysis method based on a mathematical set theory, dividing, by the processor (101), the image (300) into a plurality of sub-images (SIMG), wherein the sub-images do not overlap with each other and have the same size and each of the sub-images (SIMG) comprises a plurality of first feature points among the plurality of feature points, obtaining, by the processor, for each sub-image of the plurality sub-images a specific region (R1) and other regions in the sub-image (SIMG) via a particle swarm optimization algorithm, wherein the specific region (R1) comprises a plurality of second feature points of the first feature points and a distribution density of the second feature points in the specific region (R1) is higher than in any other regions of the sub-image, identifying, by the processor (101), a plurality of third feature points from the second feature points by selecting from the second feature points the feature points having a number of occurrences greater than a threshold value as the third feature points for performing the biometric verification operation, comprising: obtaining a plurality of optimal sets including a plurality of fourth feature points in other finger images that are for performing the biometric verification operation, and selecting, from the second feature points, the feature points having the number of occurrences greater than the threshold value as the third feature points, wherein the number of occurrences of the feature points is associated with an appearing frequency of the feature points across the second feature points and the fourth feature points, and performing, by the processor, the biometric verification operation of the user of the electronic device based on the third feature points.
  9. The method as claimed in claim 8, further comprising: determining, by the processor (101), whether the number of the plurality of third feature points is greater than a second threshold value; and when the number of the plurality of third feature points is greater than the second threshold value, selecting a plurality of fifth feature points from the third plurality of feature points by performing a sequential forward selection and performing the biometric verification operation of the user of the electronic device based on the fifth feature points, wherein the number of the plurality of fifth feature points is smaller than the number of the plurality of third feature points.
  10. The method as claimed in any of claims 8 to 9, wherein the fingerprint is characterized as a first type biological feature, the finger vein is characterized as a second type biological feature, the electronic device further comprises a storage circuit (115) comprising a plurality of secure isolated regions, and wherein the method further comprises: storing, by the processor (101), among the plurality of third feature points the feature points belonging to the first type biological feature in a second storage region of the secure isolated regions corresponding to the second type biological feature in the storage circuit (115), and storing among the plurality of third feature points the feature points belonging to the second type biological feature in a first storage region of the secure isolated regions corresponding to the first type biological feature in the storage circuit (115); wherein the electronic device further comprises a communication circuit coupled to the processor (101), and the method further comprises: obtaining a first public key, a first private key corresponding to the first public key, a second public key and a second private key corresponding to the second public key, obtaining a plurality of encrypted data received by the communication circuit comprising first encrypted data encrypted by the first public key and second encrypted data encrypted by the second public key, decrypting (S503b) the first encrypted data by the first private key and obtaining a first to-be-verified feature point of the first type biological feature, decrypting (S503a) the second encrypted data by the second private key and obtaining a second to-be-verified feature point of the second type biological feature, and performing the biometric verification operation using the first to-be-verified feature point and the second to-be-verified feature point; wherein performing the biometric verification operation using the first to-be-verified feature point and the second to-be-verified feature point comprises: obtaining (S505b) first matching data with the highest similarity from the second storage region according to the first to-be-verified feature point, obtaining (S505a) second matching data with the highest similarity from the first storage region according to the second to-be-verified feature point, determining (S509) whether an index of the first matching data is the same as an index of the second matching data, in response to the index of the first matching data being the same as the index of the second matching data, determining (S511) that the plurality of encrypted data have passed verification, and in response to the index of the first matching data not being the same as the index of the second matching data, determining (S513) that the plurality of encrypted data have not passed the verification.
  11. The method as claimed in claim 10, further comprising, before determining whether the index of the first matching data is the same as the index of the second matching data: determining (S507), by the processor (101), whether the first matching data and the second matching data meet a preset normalization standard, in response to determining that a similarity between the first matching data and the first to-be-verified feature point is greater than a threshold value, determining that the first matching data meets the preset normalization standard, and in response to determining that a similarity between the second matching data and the second to-be-verified feature point is greater than the threshold value, determining that the second matching data meets the preset normalization standard, in response to at least one of the first matching data and the second matching data not meeting the preset normalization standard, determining (S513) that the plurality of encrypted data have not passed the verification, and in response to both the first matching data and the second matching data meeting the preset normalization standard, determining (S509) whether the index of the first matching data is the same as the index of the second matching data.
  12. The method as claimed in claim 8, wherein the fingerprint is characterized as a first type biological feature, the finger vein is characterized as a second type biological feature, the electronic device further comprises a storage circuit (115), and the method further comprises: cross-combining (S803) the feature points among the plurality of third feature points that belong to the first type biological feature with the feature points among the plurality of third feature points that belong to the second type biological feature to generate combined data by the processor (101), performing (S805) a rail fence encryption algorithm on the combined data and obtaining first encrypted data, and storing (S807) the first encrypted data in the storage circuit (115), obtaining (S901a, S901b), by the processor (101), a third to-be-verified feature point of the first type biological feature and a fourth to-be-verified feature point of the second type biological feature; decrypting (S905), by the processor (101), at least one encrypted data stored in the storage circuit (115) and obtaining at least one decrypted data; comparing (S907), by the processor (101), the decrypted data with the third to-be-verified feature point and the fourth to-be-verified feature point and obtaining third matching data with the highest similarity in the first type biological feature and fourth matching data with the highest similarity in the second type biological feature; determining (S911), by the processor (101), whether the third matching data and the fourth matching data meet a preset normalization standard, wherein in response to determining that a similarity between the third matching data and the third to-be-verified feature point is greater than a threshold value, the processor (101) determines that the third matching data meets the preset normalization standard, and in response to determining that a similarity between the fourth matching data and the fourth to-be-verified feature point is greater than the threshold value, the processor (101) determines that the fourth matching data meets the preset normalization standard; in response to at least one of the third matching data and the fourth matching data not meeting the preset normalization standard, determining (S915) that the biometric verification operation is not passed; and in response to both the first matching data and the second matching data meeting the preset normalization standard, determining (S913) that the biometric verification operation is passed.
  13. The method as claimed in any of claims 8 to 12, wherein the electronic device further comprises a motor circuit (109) coupled to the processor (101), the method further comprising: receiving, by the motor circuit (109), an indication signal from the processor (101) based on the biometric verification operation of the user of the electronic device to determine whether to lock or unlock the electronic device (100).

Description

BACKGROUND Technical Field The invention relates to an electronic device and a feature obtaining method, and particularly relates to an electronic device and a method for obtaining features of biometrics. Description of Related Art Biometrics is a secure, reliable and accurate identity verification means using biological features that are usually unique, measurable, hereditary, or life-long, etc., with assistance of advanced computer technology nowadays, and can further realize applications such as automation, intelligent management, etc. At present, the most commonly used biometric techniques include fingerprint identification, human face identification, palm print identification, voice (or voiceprint) identification, vein identification, etc. The current fingerprint identification technique and finger vein identification technique are described below. [Fingerprint identification] Generally, a principle of fingerprint identification is to first perform a classification using an overall feature (for example, pattern, delta, etc.) of a fingerprint and then perform an identity recognition using a local feature (for example, position, direction, etc.). The fingerprint identification mainly includes four steps: fingerprint image reading, feature extraction, data saving and comparison. Fingerprint acquisition methods may include four types: optical, capacitive, biological radio frequency (RF), and ultrasonic. The optical type has the longest history and is most widely used, which is usually performed by putting a finger on an optical lens, and projecting the finger onto a complementary metal oxide semiconductor (CMOS) or a charge coupled device (CCD) by built-in near infrared ray (NIR) irradiation to form an image through absorption, and finally digitizing the image for processing by different fingerprint algorithms. In addition, fingerprint identification is sensitive to the state of a person being verified or the environment where the person being verified is located. For example, some people have few fingerprint features, some people's fingerprints have worn down through long use and cannot be scanned into images, some people have skin peeling on their fingers, some people's fingers are dirty, the temperature of the environment is too low, etc., and the biggest risk is that it is easy to leave traces such that people with bad intentions may be able to copy a fingerprint onto a film and use it to fool an identification device. [Finger vein identification] A principle of finger vein identification is to irradiate a finger with NIR light, so that hemoglobin flowing in veins may absorb the NIR light, and the unabsorbed NIR light may enter a sensor to obtain a clear vein image. After this image is processed by algorithms, a specific vein template is formed. It has been medically proved that everybody has a different vein image. Therefore vein template can be regarded as a unique biological feature. Finger vein identification includes four main phases: image acquisition, pre-processing, feature extraction, and feature matching. An acquisition device may be classified as a reflective type in which an NIR source and an image sensor are located on the same side, or a direct type in which an NIR source and an image sensor are respectively located on both sides. The pre-processing refers to removal of image noise. The feature extraction includes extraction of the features of lines, textures and minutiae points, and those acquired through learning. The feature matching is to compare the features with stored data. However, a defect of finger vein identification is that a finger vein may change with age and physiology, and its permanence has not been confirmed. Moreover, the equipment for acquiring a finger vein is not easy to miniaturize, a design thereof is relatively complex, and the manufacturing cost is high. Particularly, most of the related art uses a single biological feature (for example, using fingerprint only or finger vein only) for identification, and rarely mentions using multiple biological features (for example, using both fingerprint and finger vein) at the same time for identification. EP 2 833 294 A2 discloses a device and method for extracting a biometric feature vector from a fingerprint, including vein pattern information. After obtaining a biometric image, the biometric image is split into sub images (small region images) and a particular region of the biometric image containing the most characteristic features is used for biometric identification so that variability of biometric information amounts among the plurality of small region images is equal to or less than a predetermined value. US 2007/036400 A1 discloses a method for user authentication using biometric information, wherein, after obtaining a biometric image, the biometric image is split into sub images. US 2019/332755 A1 discloses a biological feature identification method. A memory is used to store a template of the biometric feature. Th