Method and apparatus for standoff detection of liveness

• US 7,417,727 B2
• Filed: 12/07/2005
• Issued: 08/26/2008
• Est. Priority Date: 12/07/2004
• Status: Expired due to Fees

• Current Assignee: Clean Earth Technologies LLC
• Original Assignee: Clean Earth Technologies LLC
• Inventors: Cilia, Andrew, Cai, Qingzhong (James), Golden, Jeffry, Polonskiy, Leonid, Boyd, Clinton, Hanebrink, Lawrence, Kaplan, Arie
• Primary Examiner(s): Evans; F. L.
• Assistant Examiner(s): Nur; Abdullahi
• Application Number:
  US11/296,804
• Publication Number:
  US 20070268485A1
• Time in Patent Office: 993 Days
• Field of Search: 356/300, 356/71, 356/317, 356/52, 356/51, 356/419, 382/124, 382/115, 600/407, 600/473, 600/475, 600/476, 600/477
• US Class Current: 356/300
• CPC Class Codes: A61B 5/0059 using light, e.g. diagnosis by transillumination, diascopy, fluorescence A61B5/0093, A61B5/1455, A61B5/02416 and A61B5/0261 take precedence
• CPC Class Codes: A61B 5/7267 involving training the classification device
• CPC Class Codes: G01N 21/31 Investigating relative effect of material at wavelengths characteristic of specific elements or molecules, e.g. atomic absorption spectrometry G01N21/72 takes precedence
• CPC Class Codes: G06V 10/143 Sensing or illuminating at different wavelengths
• CPC Class Codes: G06V 40/166 using acquisition arrangements

Abstract

A method for remotely detecting whether a subject is alive, comprising the steps of determining a calibration spectral signature for light reflectance from living skin, normalizing the calibration spectral signature values to the calibration reflectance value at a reference wavelength, storing the normalized calibration spectral signature, determining a subject spectral signature of the light reflectance of a region of skin of the subject whose liveness is to be determined, normalizing the subject spectral signature values to the subject reflectance value at the reference wavelength, comparing the normalized subject spectral signature with the normalized calibration spectral signature for at least one wavelength, generating a subject liveness signal based on the comparison of the normalized subject spectral signature with the normalized calibration spectral signature, and emitting the subject liveness signal.

Claims

1. A method for remotely detecting whether a subject is alive, comprising the steps of:

determining a calibration spectral signature for light reflectance from living skin;

normalizing the calibration spectral signature values to the calibration reflectance value at a reference wavelength;

storing the normalized calibration spectral signature;

determining a subject spectral signature of the light reflectance of a region of skin of the subject whose liveness is to be determined;

normalizing the subject spectral signature values to the subject reflectance value at the reference wavelength;

comparing the normalized subject spectral signature with the normalized calibration spectral signature for at least one wavelength;

generating a subject liveness signal based on the comparison of the normalized subject spectral signature with the normalized calibration spectral signature; and

emitting the subject liveness signal.

2. The method of claim 1 wherein determining a calibration spectral signature comprises collecting a spectral signature from a known live subject under lighting conditions that are substantially the same as those present when determining the test spectral signature.

3. The method of claim 1 wherein the light has a wavelength in a range from approximately 400 nm to approximately 700 nm.

4. The method of claim 1 wherein the reference wavelength is approximately 530 nm.

5. The method of claim 1 wherein generating a subject liveness signal further comprises generating a first signal indicating that the subject is live if the test spectral signature substantially coincides with the calibration spectral signature, and otherwise generating a second signal that the subject is not live if the test spectral signature does not substantially coincide with the calibration spectral signature.

6. The method of claim 1 wherein the at least one wavelength is at least 450 nm and less than or equal to 480 nm.

7. The method of claim 1 wherein the at least one wavelength is at least 600 nm and less than or equal to 700 nm.

8. An apparatus for remotely detecting whether a subject is alive, comprising:

a viewer and an analyzer;

wherein the viewer comprises;

an objective and a beamsplitter which together form a first intermediate image and a second intermediate image;

a reticle placed in the plane of the first intermediate image;

an eyepiece for viewing the first intermediate image through the reticle;

an optical fiber bundle comprising a plurality of sub-bundles, a first end of the optical fiber bundle being disposed in the plane of the second intermediate image such that the plurality of sub-bundles splits a central region of the second intermediate image into a plurality of light signals; and

a sensor comprising at least one photodiode assembly, wherein at least one optical fiber sub-bundle is optically coupled to the photodiode assembly;

wherein the analyzer comprises;

a digital signal processor and a memory, such that the digital signal processor generates a liveness signal if a normalized subject spectral signature is substantially equivalent to a normalized calibration spectral signature template stored in the memory and otherwise generates a not live signal.

9. The apparatus of claim 8 wherein the normalized calibration spectral signature is determined from reflectance measurements of a living human skin.

10. The apparatus of claim 8 wherein each of the plurality of light signals transmitted by the plurality of sub-bundles of optical fibers has an approximately equal intensity.

11. The apparatus of claim 8 wherein the first intermediate image comprises light that is transmitted through the beamsplitter and the second intermediate image comprises light that is reflected from the beamsplitter.

12. The apparatus of claim 11 wherein the beamsplitter reflects wavelengths of light used for detecting whether a subject is alive.

13. The apparatus of claim 11 wherein the beamsplitter transmits light having wavelengths in the range of approximately 500 nm to approximately 540 nm.

14. The apparatus of claim 8 wherein the reticle has a centrally-located reference circle for locating a region of a subject, the reference circle being optically conjugated with the first end of the optical fiber bundle such that the light transmitted to the photodiode assembly is from substantially the same region of the subject as is located within the reference circle.

15. The apparatus of claim 8 wherein the photodiode assembly comprises a lens, an optical bandpass filter, and a preamplifier, wherein the lens focuses light from an optical fiber sub-bundle through the optical bandpass filter and preamplifier onto the photodiode.

16. The apparatus of claim 15, wherein the center wavelength of the bandpass filter is optimized for the detection of difference between spectral signatures of living and dead human skins.

17. The apparatus of claim 8 wherein the optical fiber bundle is non-coherent.

18. The apparatus of claim 8 wherein the viewer further comprises a first light emitting diode optically coupled to the reticle by a first optical fiber, wherein the first light emitting diode transmits the liveness signal to the reticle.

19. The apparatus of claim 18 wherein the view further comprises a second light emitting diode optically coupled to the reticle by a second optical fiber, wherein the second light emitting diode transmits the not live signal to the reticle.

20. An apparatus for remotely detecting whether a subject is alive, comprising:

a hyperspectral imaging system comprising at least one camera for collecting a plurality of subject images formed by photons reflected from the subject, each image representing photons having wavelengths within a sub-portion of the electromagnetic spectrum;

an image processor comprising memory for storing a reference spectral signature image, and logic circuits for comparison of the subject images to the reference spectral signature image to determine whether the subject is alive, wherein the reference spectral signature image comprises an image collected from a known live sample by the hyperspectral imaging system and sent to the memory; and

an output device to indicate whether the subject is alive.

21. The apparatus of claim 20 wherein the output device comprises a display monitor for presenting an image of the subject showing sub-portions of the subject that are determined to be alive.

22. The apparatus of claim 20 wherein the output device comprises a light emitting diode.

23. The apparatus of claim 20 wherein the hyperspectral imaging system further comprises a liquid crystal tunable filter for selectively transmitting photons having wavelengths within a sub-portion of the electromagnetic spectrum to the camera.

24. The apparatus of claim 23 wherein the liquid crystal tunable filter transmits photons having wavelengths within an approximately 20 nm width band of the electromagnetic spectrum to the camera.

25. The apparatus of claim 20 wherein the hyperspectral imaging system further comprises a acousto-optical tunable filter for selectively transmitting photons having wavelengths within a sub-portion of the electromagnetic spectrum to the camera.

26. The apparatus of claim 20 wherein the comparison of the subject images to the reference spectral signature image is performed using a technique selected from the group consisting of:

least squares analysis;

spectral angle mapping; and

artificial neural network analysis.

27. A method for remotely detecting whether a subject is alive, comprising the steps of:

determining a calibration spectral signature for light reflectance from living skin;

storing the calibration spectral signature;

determining a subject spectral signature of the light reflectance of the subject whose liveness is to be determined, wherein the subject spectral signature is collected as an image comprising a plurality of pixels;

comparing the subject spectral signature with the calibration spectral signature;

generating a subject liveness signal for each pixel of the image, based on the comparison of the subject spectral signature for each pixel with the calibration spectral signature; and

displaying the subject liveness signal.

28. The method of claim 27 wherein collecting an image of the subject spectral signature comprises hyperspectral imaging.

29. The method of claim 28 wherein comparing the subject spectral signature with the calibration spectral signature comprises a technique selected from the group consisting of:

least squares analysis;

spectral angle mapping; and

artificial neural network analysis.

30. The method of claim 28 wherein hyperspectral imaging is performed using wavelengths in the range of approximately 400 nm to approximately 1700 nm.

31. The method of claim 28 wherein hyperspectral imaging is performed using wavelengths in a range of approximately 1000 nm to approximately 1600 nm.

32. The method of claim 27 wherein determining a calibration spectral signature for light reflectance from living skin comprises averaging spectral signatures from a plurality of living human skin samples.

33. The method of claim 27 further comprising:

using the subject liveness signal to identify live individuals in a crowd.

34. The method of claim 27 further comprising:

using the subject liveness signal in conjunction with another biometric indicator to determine whether the subject is wearing a disguise.

35. The method of claim 27 further comprising:

using the subject liveness signal to track a moving subject.