Apparatus and methods for capturing a fingerprint

• US 7,970,185 B2
• Filed: 06/22/2006
• Issued: 06/28/2011
• Est. Priority Date: 08/29/2005
• Status: Active Grant

• Current Assignee: Samsung Electronics Co. Ltd.
• Original Assignee: Samsung Electronics Co. Ltd., Educational Foundation Of Yonsei University (Yonsei University)
• Inventors: Kim, Sung-Jae, Lee, Chul-Han, Kim, Jai-Hie, Lee, Sang-Hoon
• Primary Examiner(s): Bali; Vikkram
• Assistant Examiner(s): Rice; Elisa M
• Application Number:
  US11/472,371
• Publication Number:
  US 20070047783A1
• Time in Patent Office: 1,832 Days
• Field of Search: 382118-127
• US Class Current: 382/124
• CPC Class Codes: G06V 40/1347 Preprocessing; Feature extraction

Abstract

An apparatus and methods for capturing a fingerprint are provided. In a first example method, first fingerprint image may be obtained in a guard region within a captured image, the guard region including less than all of the captured image. A fingerprint region may be extracted from the captured image based on a predicted color distribution, the predicted color distribution based on color information associated with the first fingerprint image, the extracted fingerprint region including the first fingerprint image within the guard region and at least a portion of the fingerprint region extending beyond the guard region within the captured image. In a second example method, guard region may be defined within a picture boundary area, the defined guard region including less than all of the picture boundary area. At least one image may be captured, the captured image spanning the picture boundary area. Information associated with the captured image may be extracted from within the guard region. Portions of the captured image associated with a user'"'"'s fingerprint may then be extracted based on the extracted information, the identified portions including a first portion within the guard region of the captured image and a second portion within the picture boundary area of the captured image outside of the guard region. In another example, an apparatus may be configured to perform either of the above-described first and second example methods.

Claims

1. A method of capturing a fingerprint, comprising:

obtaining a first fingerprint image in a guard region preset within a captured image, the guard region including less than all of the captured image; and

extracting a fingerprint region based on a predicted color distribution, the predicted color information based on color information associated with the first fingerprint image, the extracted fingerprint region including the first fingerprint image within the guard region and at least a portion of the fingerprint region extending beyond the guard region within the captured image, wherein

the extracting of the fingerprint region includes expanding the fingerprint region based on the predicted color distribution, and wherein extracting of the fingerprint region includes;

extracting color information for each pixel of the first fingerprint image;

transforming the color information into normalized color coordinates;

calculating a mean vector and a covariance matrix of the first fingerprint image based on the normalized color coordinates;

setting as target pixels, pixels which are not yet determined to be associated with one of a fingerprint region or a background region, within the captured image outside of the guard region;

extracting target color information of the target pixels and transforming the target color information into target normalized color coordinates;

calculating a distance indicating a similarity between the first fingerprint image and the target pixels based the target normalized color coordinates of the target pixels, the mean vector and the covariance matrix;

comparing the calculated distance with a distance threshold; and

expanding a temporary fingerprint region according based on a result of the comparison.

2. The method of claim 1, further comprising:

dividing the fingerprint region into a plurality of blocks;

reducing a number of outliers present within at least one of the plurality of blocks; and

extracting a ridge direction for each of the plurality of blocks.

3. The method of claim 2, further comprising:

obtaining a second fingerprint image based on the fingerprint region and the ridge direction; and

determining fingerprint characteristics by binarizing and thinning the second fingerprint image.

4. The method of claim 2, wherein reducing the number of outliers is performed based on a gradient and an initial direction of the at least one of the plurality of blocks.

5. The method of claim 3, wherein measuring the frequency is based on a at least one of discrete wavelet transform or a Tenengrad method.

6. The method of claim 2, wherein extracting the ridge direction includes:

dividing the fingerprint region into a plurality of blocks and obtaining the gradient for each of the plurality of blocks;

measuring an initial direction of each of the plurality of blocks based on the gradient;

obtaining a Euclidian norm of the gradient, for each of the plurality of blocks, by whitening the gradient and in response to the Euclidian norm;

removing one or more gradients for one or more of the plurality of blocks, the removed gradients corresponding to outliers;

remeasuring a direction of each of the plurality of blocks after the removing of the outlier gradients;

calculating a ratio of an energy sum of non-removed gradients to an energy sum of the removed gradients if a difference between the remeasured direction and a previously measured direction is greater than a threshold;

repeating the obtaining, removing, remeasuring and calculating steps if the difference between the remeasured direction and a previously measured direction is not greater than the threshold;

increasing the measuring range and then repeating the dividing, measuring, obtaining, removing, remeasuring and calculating if the ratio of the energy sums is less than an energy threshold; and

interpreting the remeasured direction as the ridge direction if the ratio is not less than the energy threshold.

7. The method of claim 6, wherein obtaining the Euclidian norm and removing the gradient further includes:

storing the energy sum of the removed gradients and the energy sum of the non-removed gradients.

8. The method of claim 1, wherein obtaining the first fingerprint image includes:

setting the guard region;

capturing an image of a finger at least within the guard region to attain the captured image;

determining whether the captured image, within the guard region, corresponds to a finger region or a background region; and

interpreting the captured image, within the guard region, as the first fingerprint image if the captured image, within the guard region, is determined to correspond to the finger region.

9. The method of claim 8, wherein the determining of whether the captured image within the guard region corresponds to a finger region or a background region includes:

measuring a frequency of the captured image within the guard region;

comparing the measured frequency with a frequency threshold;

determining the captured image, within the guard region, to correspond to a finger region if the measured frequency is greater than the frequency threshold; and

determining the captured image, within the guard region, to correspond to a background region if the measured frequency is not greater than the frequency threshold.

10. The method of claim 9, further comprising:

repeating at least the obtaining and extracting steps if the captured image, within the guard region, is determined to correspond with the background region.

11. The method of claim 1, wherein at least one of the normalized color coordinates (r_y, g_y, b_y) and target normalized color coordinates (r_y, g_y, b_y) are obtained by the following equations:

12. The method of claim 1, wherein the mean vector and covariance matrix are obtained based on at least two elements associated with the normalized color coordinates.

13. The method of claim 11, wherein the mean vector (m) and covariance matrix (Σ

) are calculated based on the following equations;

14. The method of claim 1, wherein the temporary fingerprint region becomes the extracted fingerprint region when no additional expansions are possible.

15. The method of claim 1, wherein expanding the temporary fingerprint region includes

determining that the target pixel corresponds to a fingerprint region if the distance is less than the distance threshold,;

and

determining the target pixel corresponds to a background region if the distance is greater than the distance threshold.

16. The method of claim 1, wherein the distance (D(i)) is a Mahalanobis distance expressed as follows:

D(i)=(x(i)−

m)^TΣ

<sup>−

1</sup>(x(i)−

m)

wherein x(i) denotes the normalized color coordinates of the target pixel, m is the calculated mean vector, and Σ

<sup>−

1</sup> is the calculated covariance.

17. An apparatus performing the method of claim 1.

18. An apparatus for capturing a fingerprint, comprising:

a fingerprint image acquisition unit obtaining a first fingerprint image in a guard region preset within a captured image, the guard region including less than all of the captured image; and

a fingerprint region extraction unit extracting a fingerprint region based on a predicted color distribution, the predicted color information based on color information associated with the first fingerprint image, the extracted fingerprint region including the first fingerprint image within the guard region and at least a portion of the fingerprint region extending beyond the guard region within the captured image,

wherein

the extraction of the fingerprint region by the fingerprint region extraction unit includes expanding the fingerprint region based on the predicted color distribution, and wherein the fingerprint region extraction unit includes

a color information calculation unit extracting color information for each pixel of the first fingerprint image, transforming the color information into normalized color coordinates and calculating a mean vector and a covariance matrix of the first fingerprint image based on the normalized color coordinates;

a target pixel storage unit setting as target pixels, pixels which are not yet determined to be associated with one of a fingerprint region or a background region, within the captured image outside of the guard region and extracting target color information of the target pixels and transforming the target color information into target normalized color coordinates;

a distance calculation unit calculating a distance indicating a similarity between the first fingerprint image and the target pixels based the target normalized color coordinates of the target pixels, the mean vector and the covariance matrix;

a fingerprint region expansion unit comparing the calculated distance with a distance threshold and expanding a temporary fingerprint region according based on a result of the comparison.

19. The apparatus of claim 18, further comprising:

a ridge direction extraction unit dividing the fingerprint region into a plurality of blocks, reducing a number of outliers present within at least one of the plurality of blocks and extracting a ridge direction for each of the plurality of blocks;

a fingerprint characteristic extraction unit obtaining a second fingerprint image based on the fingerprint region and the ridge direction and determining fingerprint characteristics by binarizing and thinning the second fingerprint image.

20. The apparatus of claim 18, wherein the fingerprint image acquisition unit includes:

a guard region setting unit setting the guard region; and

an image acquisition unit capturing an image of a finger at least within the guard region to attain the captured image, determining whether the captured image, within the guard region, corresponds to a finger region or a background region and interpreting the captured image, within the guard region, as the first fingerprint image if the captured image, within the guard region, is determined to correspond to the finger region.

21. The apparatus of claim 20, wherein the image acquisition unit measures a frequency of the captured image within the guard region, compares the measured frequency with a frequency threshold, determines the captured image, within the guard region, to correspond to a finger region if the measured frequency is greater than the frequency threshold and determines the captured image, within the guard region, to correspond to a background region if the measured frequency is not greater than the frequency threshold.

22. The apparatus of claim 21, wherein the frequency is measured based on at least one of a discrete wavelet transform and a Tenengrad method.

23. The apparatus of claim 18, wherein at least one of the normalized color coordinates (r_y, g_y, b_y) and target normalized color coordinates (r_y, g_y, b_y) are obtained by the following equation:

24. The apparatus of claim 18, wherein the mean vector and covariance matrix are obtained based on at least two elements associated with the normalized color coordinates.

25. The apparatus of claim 18, wherein the mean vector (m) and covariance matrix (Σ

) are calculated based on the following equation;

26. The apparatus of claim 18, wherein the temporary fingerprint region becomes the extracted fingerprint region when no additional expansions are possible.

27. The apparatus of claim 18, wherein the fingerprint region expansion unit determines that the target pixel corresponds to a fingerprint region if the distance is less than the distance threshold and determines the target pixel corresponds to a background region if the distance is greater than the distance threshold.

28. The apparatus of claim 18, wherein the distance (D(i)) is a Mahalanobis distance expressed as follows:

D(i)=(x(i)−

m)^TΣ

<sup>−

1</sup>(x(i)−

m)

wherein x(i) denotes the normalized color coordinates of the target pixel, m is the calculated mean vector, and Σ

<sup>−

1</sup> is the calculated covariance.

29. The apparatus of claim 19, wherein the ridge direction extraction unit includes:

an initial direction measuring unit dividing the fingerprint region into a plurality of blocks, obtaining the gradient for each of the plurality of blocks and measuring an initial direction of each of the plurality of blocks based on the gradient;

an outlier removal unit obtaining a Euclidian norm of the gradient, for each of the plurality of blocks, by whitening the gradient and in response to the Euclidian norm, removing gradients corresponding to outliers, and storing an energy sum of the removed gradients and an energy sum of non-removed gradients;

a direction remeasuring a direction of each of the plurality of blocks after the removing of the outlier gradients;

an energy calculation unit calculating a ratio of an energy sum of non-removed gradients to an energy sum of the removed gradients if a difference between the remeasured direction and a previously measured direction is greater than a threshold;

a ridge direction determination unit interpreting the remeasured direction as the ridge direction if the ratio is not less than the energy threshold.