Signature is one of the most popular and commonly accepted biometric hallmarks that has been used since the ancient times for verifying different entities related to human beings, viz. documents, forms, bank checks, individuals, etc. Therefore, signature verification is a critical task and many efforts have been made to remove the uncertainty involved in the manual authentication procedure, which makes signature verification an important research line in the field of machine learning and pattern recognition. In this notebook, we model a writer independent signature verification task with a convolutional Siamese network.
The BHSig260 signature dataset contains the signatures of 260 persons, among them 100 were signed in Bengali and 160 are signed in Hindi.
For each of the signers, 24 genuine and 30 forged signatures are available. This results in 100 × 24 = 2, 400 genuine and 100 × 30 = 3, 000 forged signatures in Bengali, and 160 × 24 = 3, 840 genuine and 160×30 = 4, 800 forged signatures in Hindi.
In this task we are considering only Hindi singatures for easeness.
Paper Link: https://arxiv.org/pdf/1707.02131.pdf