Efficient deep learning model for signature verification
DeepSign is an efficient deep learning model for signature verification. Signatures are a common form of verification and a target for fraud. Current state of the art verification models involve siamese convolutional neural networks. We similarly apply siamese convolutional neural networks that are both efficient and high performing for the task of signature verification. We created our own SqueezeNet-inspired efficient architecture, DeepSign, that uses 65% fewer parameters than Google’s MobileNetv2 and 97% fewer parameters than the current state of the art, SigNet. This lightweight model is readily applicable to mobile devices for both online or offline signature verification.
DeepSign is integrated into a React web app so that viewers can demo and test their own signatures against the model. To view both the live web demo and our academic paper, visit deepsign.ml.