Method of detecting fraud of an iris recognition system

A method of detecting attempted fraud against a recognition system for recognising the iris of the human eye, comprising:

• generation of a first image of an iris of the human eye observed using first means of image capture sensitive in a visible light spectrum and over a frequency range covering at least 50% of the visible light spectrum, wherein this first image is a broad spectrum colour image;
• generation of a second image of said iris observed using second means of image capture sensitive in a near infra-red spectrum;
• determination of at least one characteristic of the first image determined as a function of respective optical characteristics of pixels of a plurality of pixels of this first image;
• and
• determination of at least one characteristic of the second image determined as a function of respective luminous intensities of pixels of a plurality of pixels of this second image;
• and
• as a function of said at least one characteristic of the first image thus determined and of said at least one characteristic of the second image thus determined, a signal representative of suspected detection of attempted fraud will be generated.