Automatic Offline Signature Verification

Offline signature verification is a biometric verification task that aims to discriminate between genuine and forged samples of handwritten signatures. This is a particularly important form of verification due to the ubiquitous use of handwritten signatures as a means of personal identification in legal contracts, administrative forms, and financial documents.

In this series of work, we frame signature verification as a deep metric learning problem, where we ultimately learn a model that yields semantically meaningful representations of signatures such that vectors for similar signatures are close and dissimilar signatures are far apart given a distance metric.

We experiment with different metric learning objectives (contrastive & triplet loss), as well as various model architectures, and document our findings in a three-part blog series.