Une approche de type comprehension de scene appliquee au probleme de la verification automatique de l'identite par l'image de la signature manuscrite

In the light of previous works in the field of Automatic Handwritten Signature Verification Systems (AHSVS), an image-understanding system based on the extraction of a novel signature representation from gray-level images is proposed for automatic signature verification.; By an image-understanding system we mean the separation of the local and the global evaluation functions for the analysis of the handwritten signature. The approach considered in this thesis is text insensitive because no attempt is made to segment specific letters from the semantic part of the signature. Two classes of features, static and pseudo-dynamic, are taken into account for the representation of the scene. The former is related to the geometric shape and the spatial relations between some reference and test primitives extracted from the signature line; this constitutes the originality of the proposed strategies applied to the AHSVS problem in this thesis. The latter is associated with the gray-level evaluated inside the primitives located on the static solution path. The proposed AHSVS is able to compare two structural graphs related to the corresponding representation (primitives sets) of the reference and test handwritten signature images, permitting the implementation of characteristics related both to static and pseudo-dynamic features evaluated locally along the signature line.; The automatic handwritten signature verification system is used with a handwritten signature database consisting of 248 handwritten signature images from eight writers, 224 random forgeries, 160 tracing forgeries from one forger using two kinds of tracing pens and 224 photocopies of the genuine specimens. This database is used for evaluating the capacity of the proposed AHSVS to discriminate between individuals (random forgeries) and the capacity to distinguish between genuines and forgeries produced with or without imitation (of the genuine).; Practically all tracings and photocopies are accepted by the system in the static situation.; Using both the static and the pseudo-dynamic schemes allow the partial elimination of skilled forgeries characterized by a high similarity in the static form, and showing marked gray-level dissimilarity along the signature line. (Abstract shortened by UMI.)...