CapsNet Regularization and its Conjugation with ResNet for Signature Identification

Mahdi Jampour, Saeid Abbaasi and Malihe Javidi

Congrats and thanks to Dr. Javidi and Mr Abbasi for their collaboration on our recently accepted paper with the title of "CapsNet Regularization and its Conjugation with ResNet for Signature Identification" in prestigious peer-reviewed journal of Pattern Recognition. The PR is a journal in top 5% in Artificial Intelligence with impact factor 7.74.

 

Abstract

We propose a new regularization term for CapsNet that significantly improves the generalization power of the original method from small training data while requiring much fewer parameters, making it suitable for large input images. We also propose a very efficient DNN architecture that integrates CapsNet with ResNet to obtain the advantages of the two architectures. CapsNet allows a powerful understanding of the objects' components and their positions, while ResNet provides efficient feature extraction and description. Our approach is general, and we demonstrate it on the problem of signature identification from images. To show our approach superiority, we provide several evaluations with different protocols. We also show that our approach outperforms the state-of-the-art on this problem  with thorough experiments on three publicly available datasets CEDAR, MCYT, and UTSig.

 

Link to the paper on ScienceDirect:     https://doi.org/10.1016/j.patcog.2021.107851

Preprint is available on ResearchGate: https://www.researchgate.net/publication/348871709

Link to the source code on GitHub:      https://github.com/Javidi31/RegCapsNet

Paper is open access from the following link till August 29, 2021:
https://authors.elsevier.com/a/1dNsQ77nKcuF4

 

Mahdi Jampour, Saeid Abbaasi, Malihe Javidi,
CapsNet Regularization and its Conjugation with ResNet for Signature Identification,
Pattern Recognition, Volume 120, 2021, 107851, https://doi.org/10.1016/j.patcog.2021.107851.