Method for detecting the reality of venous networks for the purposes of identifying individuals, and biometric recognition method

1. A method of detecting a venous network in a portion of a human body, the method comprising the steps of:

capturing an infrared image of said human body portion in order to reveal a venous network of said human body portion;

analyzing contrast in the infrared image in order to determine contrast characteristics of the infrared image; and

determining whether the human body portion is real on the basis of contrast characteristics determined in the infrared image and of contrast characteristics determined from reference human body portions, wherein

the method comprises a configuration stage and a detection stage, the configuration stage comprising the steps of:

capturing infrared images of human body portions and representations of venous networks;

isolating blocks of pixels in each image and performing contrast analysis to determine for each block a directional gradient;

determining average energies for the directional gradients; and

determining by training average energy thresholds for discriminating between an image of a venous network captured directly from a human body portion and an image of a venous network captured from a representation of a venous network;

the detection stage comprising the steps of:

capturing at least one infrared image of a venous network;

isolating blocks of pixels in the image and performing contrast analysis to determine for each block a directional gradient;

determining average energies of the directional gradients; and

comparing these average energies with the thresholds in order to verify that the image of the venous network was captured directly from a human body portion.

2. The method according to claim 1 , wherein the directional gradients are determined for at least two image resolutions.

3. The method according to claim 1 , comprising the steps of:

isolating blocks of a first number of pixels in each image and performing contrast analysis in each of the blocks in order to determine a directional gradient in each block of the first number of pixels;

isolating blocks of a second number of pixels greater than the first number of pixels in each image; and

performing contrast analysis in each of the blocks in order to determine a directional gradient in each block of the second number of pixels.

4. The method according to claim 3 , wherein the second number of pixels is equal to the square of the first number of pixels.

5. The method according to claim 4 , wherein the first number of pixels is equal to nine.

6. The method according to claim 1 , wherein the blocks are superposed so as to cover all of the images.

7. The method according to claim 1 , wherein:

the vectors are determined from average energies of directional gradients;

during the configuration stage, an SVM type training machine is used for determining a separation surface for the average energy vectors of the directional gradients corresponding to images of venous networks captured directly from a human body portion and average energy vectors of directional gradients of venous network images captured on representations of venous networks; and

during the detection stage, the mean energy vectors of the captured image are projected onto the separation surface so as to determine as a function of the sign of the projection whether the image was captured directly from a human body portion or from a representation of a venous network.

8. The method according to claim 1 , wherein the directional gradients are determined by:

subtracting gray levels of the pixels; or

Sobel type filtering; or

Canny Deriche type filtering.

9. A biometric recognition method comprising a stage of detecting whether a human body portion is real, and a biometric recognition stage, the method comprising the following steps:

capturing biometric characteristics on a portion of an individual's body;

comparing the biometric characteristics with stored biometric characteristics; and

on the basis of the comparison determining whether or not the individual is recognized;

the detection stage being performed by implementing the method according to claim 1 , and the recognition stage being performed only if the human body portion is determined as being real at the end of the detection stage.