Magnetic detection of replay attack

1. A method of detecting a replay attack on a voice biometrics system, the method comprising:

receiving an audio signal representing speech;

detecting a magnetic field;

determining if there is a correlation between the audio signal and the magnetic field, wherein determining if there is a correlation between the audio signal and the magnetic field comprises:

sampling the detected magnetic field at a first sample rate;

sampling the audio signal at a second sample rate; and

determining if there is a correlation between the sampled audio signal and the sampled detected magnetic field; and

if there is a correlation between the audio signal and the magnetic field, determining that the audio signal may result from a replay attack.

2. A method according to claim 1 , wherein determining if there is a correlation between the audio signal and the magnetic field comprises:

identifying first periods during which the audio signal contains speech;

identifying second periods during which the magnetic field differs from a baseline; and

determining if the first and second periods are substantially the same.

3. A method according to claim 2 , comprising determining that the first and second periods are substantially the same if more than 60% of the first periods during which the audio signal contains speech overlap with the second periods during which the magnetic field differs significantly from a baseline, and/or more than 60% of the second periods during which the magnetic field differs significantly from a baseline overlap with first periods during which the audio signal contains speech.

4. A method according to claim 3 , comprising determining that the first and second periods are substantially the same if more than 80% of the first periods during which the audio signal contains speech overlap with second periods during which the magnetic field differs significantly from a baseline, and/or more than 80% of the second periods during which the magnetic field differs significantly from a baseline overlap with first periods during which the audio signal contains speech.

5. A method according to claim 1 , comprising:

receiving a series of values of a signal representing a magnetic field strength;

forming an average value of the magnetic field strength over a period of time; and

subtracting the average value of the magnetic field strength from the series of values of the signal representing the magnetic field strength to form said detected magnetic field.

6. A method according to claim 1 , comprising:

obtaining a digital audio signal at a third sample rate, and undersampling said digital audio signal to form said audio signal at said second sample rate.

7. A method according to claim 1 , wherein said second sample rate is approximately equal to said sample rate.

8. A method according to claim 1 , wherein the step of determining if there is a correlation between the sampled audio signal and the sampled detected magnetic field comprises performing a mathematical correlation operation on the sampled audio signal and the sampled detected magnetic field to obtain an output correlation function, and determining if a peak value of the output correlation function exceeds a predetermined threshold.

9. A method according to claim 1 , further comprising:

determining a direction of a source of said audio signal representing speech;

determining a direction of a source of said magnetic field; and

determining that the audio signal may result from a replay attack if the direction of the source of said audio signal representing speech corresponds to the direction of the source of said magnetic field.

10. A system for detecting a replay attack on a voice biometrics system, the system comprising:

an input for receiving an audio signal representing speech;

a detector for detecting a magnetic field; and

a processor for:

determining if there is a correlation between the audio signal and the magnetic field, wherein determining if there is a correlation between the audio signal and the magnetic field comprises:

sampling the detected magnetic field at a first sample rate;

sampling the audio signal at a second sample rate; and

determining if there is a correlation between the sampled audio signal and the sampled detected magnetic field; and

if there is a correlation between the audio signal and the magnetic field, determining that the audio signal may result from a replay attack.

11. A computer program product, comprising a non-transitory computer-readable tangible medium, and instructions for performing a method according to claim 1 .