Process for marking a multimedia document, such as an image, by generating a mark

Method of marking a multi-media document lorig by generating a mark from said document and a marking image Iinf_0, comprising the following steps: decomposing the initial document lorig according to a lossy approximation method ; determining a reference image lref ; point-by-point subtracting the initial document lorig and the reference image lref., and according to at least one vector representative of the image in order to produce a support image Is ; formatting the marking image Iinf_0 so as to obtain an image Iinf1 of the same size as the support image Is by pre-introducing at least one level of redundancy; point-by-point combining the marking image Iinf_0, formatted beforehand, with the support image so as to generate a mark or a signature which can be, according to the case, either filed, or internally reintroduced into the initial document lorig. The combining of image Iinf1 and support image Is produces a compact-sized mark which can be either stored or reintegrated within the document itself.