Systems and methods for performing fingerprint based user authentication using imagery captured using mobile devices

• US 9,361,507 B1
• Filed: 08/06/2015
• Issued: 06/07/2016
• Est. Priority Date: 02/06/2015
• Status: Active Grant

• Current Assignee: Veridium IP Limited (Harvard Investments LP)
• Original Assignee: Hoyos Labs IP, Ltd. (Harvard Investments LP)
• Inventors: Othman, Asem, Tyson, Richard, Simpson, Andrew, Hoyos, Hector, Mather, Jonathan Francis
• Primary Examiner(s): Strege, John
• Application Number:
  US14/819,639
• Time in Patent Office: 306 Days
• Field of Search: 382/124
• US Class Current: 1/1
• CPC Class Codes: G06V 10/141 Control of illumination
• CPC Class Codes: G06V 10/255 Detecting or recognising potential candidate objects based on visual cues, e.g. shapes
• CPC Class Codes: G06V 10/993 Evaluation of the quality of the acquired pattern
• CPC Class Codes: G06V 30/142 using hand-held instruments; Constructional details of the instruments
• CPC Class Codes: G06V 40/1312 direct reading, e.g. contactless acquisition
• CPC Class Codes: G06V 40/1353 Extracting features related to minutiae or pores
• CPC Class Codes: G06V 40/1388 using image processing
• CPC Class Codes: G06V 40/1394 using acquisition arrangements

Abstract

Technologies are presented herein in support of a system and method for performing fingerprint recognition. Embodiments of the present invention concern a system and method for capturing a user'"'"'s biometric features and generating an identifier characterizing the user'"'"'s biometric features using a mobile device such as a smartphone. The biometric identifier is generated using imagery captured of a plurality of fingers of a user for the purposes of authenticating/identifying the user according to the captured biometrics and determining the user'"'"'s liveness. The present disclosure also describes additional techniques for preventing erroneous authentication caused by spoofing. In some examples, the anti-spoofing techniques may include capturing one or more images of a user'"'"'s fingers and analyzing the captured images for indications of liveness.

Claims

1. A method for performing fingerprint recognition, the method comprising:

capturing, by a mobile device having a camera, a storage medium, instructions stored on the storage medium, and a processor configured by executing the instructions, one or more high-resolution images, wherein each of the one or more high-resolution images depict a plurality of fingers of a subject;

identifying, with the processor from the one or more high-resolution images according to a segmentation algorithm, a respective fingertip segment for each finger among the plurality of fingers, wherein the identifying step includes;

analyzing, with the processor, at least a portion of the one or more high-resolution images that depicts the plurality of fingers to identify a plurality of candidate fingertip segments;

filtering, with the processor, the plurality of candidate fingertip segments according to measured physical characteristics and expected physical characteristics relating to one or more of the plurality of fingers and individual fingers, wherein the step of filtering the plurality of candidate fingertip segments comprises;

identifying one or more sets of candidate fingertip segments, for each of the one or more sets;

measuring physical characteristics of the candidate fingertip segments in the set,

comparing the measured physical characteristics of the candidate fingertip segments in the set to corresponding expected physical characteristics,

calculating a weighted score for each candidate fingertip segment in the set according to the comparison, wherein each comparison has an associated weight defined as a function of reliability,

calculating a score for the set according to the weighted scores, and

selecting a best set according to the calculated score;

extracting, with the processor for each candidate fingertip segment in the best set, discriminatory features from the respective candidate fingertip segments; and

generating a biometric identifier including the extracted discriminatory features;

storing the generated biometric identifier in the memory with the processor.

2. The method of claim 1, further comprising:

capturing, with the processor using the camera, one or more initial images depicting the plurality of fingers;

detecting, with the processor using a finger finding algorithm, the plurality of fingers depicted in the one or more initial images; and

determining, with the processor from the one or more initial images, that the plurality of fingers are in-focus.

3. The method of claim 2, further comprising:

triggering, with the processor, the camera to capture the one or more high-resolution images in response to detecting the plurality of fingers in the one or more initial images and determining that the plurality of fingers are in-focus.

4. The method of claim 1, wherein the step of identifying a respective fingertip segment for each finger comprises:

distinguishing, with the processor using a major classifier, a major region within the one or more high-resolution images that includes the plurality of fingers from a background region; and

identifying, with the processor within the major region;

a respective fingertip segment for each finger.

5. The method of claim 1, wherein the step of identifying the respective fingertip segment for each finger further comprises:

analyzing, with the processor, at least a portion of the one or more high-resolution images that depicts the plurality of fingers using one or more minor classifiers to identify the plurality of candidate fingertip segments.

6. The method of claim 1, wherein the measured physical characteristics include one or more of:

relative width of one or more candidate fingertip segments;

relative lengths of the candidate fingertip segments in the set;

relative positions in a horizontal direction of the candidate fingertip segments in the set; and

relative positions in a vertical direction of the candidate fingertip segments in the set.

7. The method of claim 6, further comprising:

determining a total width of the plurality of fingers; and

wherein the expected physical characteristics are a function of the total width.

8. The method of claim 4, further comprising:

segmenting, with the processor, at least a portion of the one or more high-resolution images that depicts the plurality of fingers into a plurality of minor regions;

classifying, with the processor implementing one or more minor classifiers, a subset of minor regions as fingertip segments; and

identifying among the subset of minor regions, a respective fingertip segment for each of the fingers.

9. The method of claim 8, wherein the plurality of minor classifiers are implemented in conjunction with additional finger-tip location finding algorithms.

10. The method of claim 8, further comprising:

wherein the respective fingertip segment for each of the fingers are identified according to expected physical characteristics of the plurality of fingers.

11. The method of claim 8, further comprising, enhancing the minor regions depicted in the one or more high-resolution images prior to the step of extracting the discriminatory features of the respective fingertip segments.

12. The method of claim 11, wherein enhancing the minor regions comprises:

generating, with the processor, a smoothed ridge orientation map for each of the minor regions; and

enhancing each of the smoothed ridge orientation maps by one or more of;

applying a band pass filter that is configured to pass fingerprint ridges frequencies and minimize background image frequencies, and

performing a histogram equalization of pixel intensities, wherein the histogram equalization includes plotting a histogram of pixel intensities, and adjusting the intensities such that the histogram includes approximately equal amounts of each intensity across the histogram.

13. The method of claim 1, wherein extracting discriminatory features for each of the respective fingertip segments comprises:

extracting from each of the respective fingertip segments, minutiae of a respective finger according to a minutiae extracting algorithm; and

wherein generating the biometric identifier comprises, for each of the fingers, encoding feature vectors representing the extracted minutia in the memory.

14. The method of claim 1, further comprising:

matching, with the processor, the generated biometric identifier to a previously stored biometric identifier associated with the user;

generating, based on the matching, a similarity score between the generated biometric identifier and the stored biometric identifier; and

verifying that the subject is the user based on the similarity score.

15. The method of claim 14, further comprising:

determining, with the processor, that the generated biometric identifier is representative of a live subject.

16. The method of claim 1, further comprising, prior to the extracting step:

constructing, with the processor in the memory, for each of the plurality of fingers, a border around the respective fingertip segment to highlight the respective fingertip segment;

displaying, with the processor on a display of the mobile device, the one or more high resolution images and the borders; and

prompting, with the processor on a display of the mobile device, the user to confirm the locations of the borders highlighting the fingertip segments.

17. A method for performing fingerprint recognition, the method comprising:

capturing, by a mobile device having a camera, a storage medium, instructions stored on the storage medium, a display, and a processor configured by executing the instructions, one or more initial images depicting a finger of a subject,

wherein each of the one or more high-resolution images depicts a finger of the subject;

identifying, with the processor from the one or more high-resolution images according to a segmentation algorithm, a respective fingertip segment for the finger, wherein the identifying step includes;

analyzing, with the processor, at least a portion of the one or more high-resolution images to identify a plurality of candidate fingertip segments,

filtering, with the processor, the plurality of candidate fingertip segments according to measured physical characteristics and expected physical characteristics relating to the finger, wherein the step of filtering includes;

for each candidate fingertip segment,

measuring physical characteristics of the candidate fingertip segment,

comparing the measured physical characteristics of the candidate fingertip segment to one or more corresponding expected physical characteristics,

calculating a weighted score for the candidate fingertip segment according to the comparison of the physical characteristics, wherein each comparison has an associated weight defined as a function of reliability, and

selecting the respective fingertip segment from among the plurality of candidate fingertip segments according to the calculated weighted scores;

extracting, with the processor from the respective fingertip segment, discriminatory features of the finger;

generating a biometric identifier including the extracted discriminatory features; and

storing the generated biometric identifier in the memory with the processor.

18. The method of claim 17, wherein the step of identifying the respective fingertip segment comprises:

distinguishing, with the processor using a major classifier, a major region within the one or more high-resolution images that depicts the finger from a background region; and

identifying, with the processor within the major region;

the respective fingertip segment for the finger.

19. The method of claim 17, wherein the step of identifying the respective fingertip segment further comprises:

analyzing, with the processor, the at least a portion of the one or more high-resolution images that depicts the finger using one or more minor classifiers to identify the plurality of candidate fingertip segments.

20. The method of claim 17, wherein the step of identifying the respective fingertip segment further comprises:

segmenting, with the processor, at least a portion of the one or more high-resolution images that depicts the finger into a plurality of minor regions; and

classifying, with the processor implementing one or more minor classifiers, a subset of minor regions as fingertip segments.

21. The method of claim 20, wherein the one or more minor classifiers are implemented in conjunction with additional finger-tip location finding algorithms.

22. The method of claim 17, further comprising:

capturing, by the processor using the camera, one or more initial images depicting the finger of the subject;

displaying the one or more initial images in near-real time on the display of the mobile device;

displaying, with the processor on the display, a finger placement guide, wherein the finger placement guide is overlaid the one or more initial images and prompts the user to align the finger in the initial image with the guide, wherein a finger is in alignment with the guide when the finger is at a particular location in the camera'"'"'s field of view and at a particular distance from the camera;

detecting, with the processor using a finger finding algorithm, the finger depicted in the one or more initial images; and

determining, with the processor from the one or more initial images, that the finger depicted in the one or more initial images is in alignment within the guide and is in-focus;

triggering, with the processor, the camera to capture the one or more high-resolution images in response to detecting the plurality of fingers in the one or more initial images and determining that the plurality of fingers are in alignment with the guide and in-focus.

23. The method of claim 20, further comprising, enhancing the identified respective fingertip segment prior to the step of extracting the discriminatory features of the respective fingertip segments.

24. The method of claim 23, wherein enhancing the respective fingertip segment comprises:

generating, with the processor, a smoothed ridge orientation map for the respective fingertip segment; and

enhancing each of the smoothed ridge orientation maps by one or more of;

applying a band pass filter that is configured to pass fingerprint ridges frequencies and minimize background image frequencies, and

performing a histogram equalization of pixel intensities, wherein the histogram equalization includes plotting a histogram of pixel intensities, and adjusting the intensities such that the histogram includes approximately equal amounts of each intensity across the histogram.

25. The method of claim 17, wherein extracting discriminatory features for the respective fingertip segment comprises:

extracting from the respective fingertip segment, minutiae of the finger according to a minutiae extracting algorithm; and

wherein generating the biometric identifier comprises encoding feature vectors representing the extracted minutia in the memory.

26. The method of claim 17, further comprising:

matching, with the processor, the generated biometric identifier to a previously stored biometric identifier associated with the user;

generating, based on the matching, a similarity score between the generated biometric identifier and the stored biometric identifier;

verifying that the subject is the user based on the similarity score; and

determining, with the processor, that the generated biometric identifier is representative of a live subject.

27. The method of claim 17, further comprising, prior to the extracting step:

constructing, with the processor in the memory, a border around the respective fingertip segment to highlight the respective fingertip segment;

displaying, with the processor on the display of the mobile device, the one or more high resolution images and the border; and

prompting, with the processor on the display of the mobile device, the user to confirm the locations of the border highlighting the fingertip segment.