Method and an apparatus for determining vein patterns from a colour image
The present invention is directed to a method of determining vein patterns from a color image for personal identification, the method comprising forming a counterpart of the color image by applying a functional relationship obtained from optimization on the color image, wherein the counterpart of the color image comprises the vein patterns. An apparatus for determining vein patterns from a color image is also disclosed.
1. A method of determining vein patterns from a colour image for personal identification, the method comprising:
forming a counterpart of the colour image by applying a functional relationship obtained from optimization on the colour image, wherein the counterpart of the colour image comprises the vein patterns,
wherein the functional relationship is formulated using a model modelling colour image and a counterpart of the modelling colour image, wherein the counterpart of the modelling colour image comprises modelling vein patterns.
2. The method as claimed in claim 1 , wherein in forming the counterpart of the colour image, applying the functional relationship on the colour image is applying a neural network on the colour image.
3. The method as claimed in claim 2 , wherein forming the counterpart of the colour image further comprises training the neural network by back-propagation.
4. The method as claimed in claim 1 , wherein forming the counterpart of the colour image further comprises selecting the functional relationship from a plurality of functional relationships between modelling colour images and counterparts of the modelling colour images.
5. The method as claimed in claim 4 , further comprising grouping parts of the counterpart of the colour image.
6. The method as claimed in claim 5 , wherein grouping parts of the counterpart of the colour image comprises assigning pixels in the IR image having substantially the same colour to a same group.
7. The method as claimed in claim 1 , wherein the counterpart of the modelling colour image comprises an infrared (IR) image or a near-infrared (NIR) image.
8. The method as claimed in claim 7 , wherein the functional relationship comprises a plurality of functional relationships between colour images and IR images wherein each colour image and the respective IR image thereof forms an image pair; and wherein for at least one of the image pairs, the colour image and the respective IR image are captured simultaneously.
9. The method as claimed in claim 8 , wherein forming the counterpart of the colour image further comprises training with a set of image pairs formed by the modelling colour images and the IR images thereof.
10. The method as claimed in claim 7 , wherein the functional relationship comprises a parametric vector for predicating the IR image from the colour image.
11. The method as claimed in claim 1 , further comprising removing blocking artifacts in the colour image, wherein the colour image is a compressed colour image.
12. The method as claimed in claim 11 , wherein removing blocking artifacts comprises using a knowledge-based approach or a one-pass algorithm.
13. The method as claimed in claim 1 , further comprising extracting vein patterns from an image, wherein the image comprises a plurality of pixels.
14. The method as claimed in claim 1 , wherein the colour image is a true colour image.
15. A method of determining vein patterns from a colour image for personal identification, the method comprising:
forming a counterpart of the colour image by applying a functional relationship obtained from optimization on the colour image, wherein the counterpart of the colour image comprises the vein patterns,
wherein the functional relationship obtained from optimization is formulated using an approximate regression comprising a least-squares-error energy function with an optimal weight.
16. A method of determining vein patterns from a colour image for personal identification, the method comprising:
forming a counterpart of the colour image by applying a functional relationship obtained from optimization on the colour image, wherein the counterpart of the colour image comprises the vein patterns,
wherein the colour image comprises colour components;
wherein the counterpart of the colour image comprises skin-characterizing variables; and
wherein forming a counterpart of the colour image comprises determining distributions of the skin-characterizing variables to form the vein patterns.
17. The method as claimed in claim 16 , wherein determining distributions of the skin-characterizing variables comprises applying an inverse of the functional relationship on the colour components obtained from the colour image.
18. The method as claimed in claim 17 , wherein the colour components comprise a red (R) component, a green (G) component, and a blue (B) component.
19. An apparatus for determining vein patterns from a colour image for personal identification, the apparatus comprising:
a determining unit configured to form a counterpart of the colour image by applying a functional relationship obtained from optimization on the colour image, wherein the counterpart of the colour image comprises the vein patterns,
wherein the functional relationship is formulated using a modelling colour image and a counterpart of the modelling colour image, wherein the counterpart of the modelling colour image comprises modelling vein patterns.
20. The apparatus as claimed in claim 19 , wherein the colour image is a true colour image.