Graph-based Signature Verification

Most approaches for automatic signature verification rely on statistical pattern recognition using fixed-size feature vectors. These forms of representation have known limitations including the inability to capture the global structure of the signatures and the relations between their subparts in a standardized way. These limitations can be overcome using graph-based representations and structural pattern recognition. Graphs offer a more intuitive, powerful, and flexible mechanism for formally representing and comparing handwritten signatures. However, graph-based pattern recognition has a high computational complexity involved in structural matching and requires efficient approximations and algorithms to be used for a real-world application like signature verification.

In this research project, we investigate the usefulness of graph-based methods for the task of offline signature verification. We have published several research papers and articles which highlight the advantages and disadvantages of structural approaches for offline signature verification.