/ConvArc

PyTorch implementation of Attentive Recurrent Comparators

Primary LanguagePythonMIT LicenseMIT

ConvArc

This repo provides a Pytorch implementation for Recurrent Comparator with attention models to detect counterfeit documents.

Cite

@article{aberenguel_recurrent_counterfeit,
  author    = { Albert Berenguel and
                Oriol Ramos Terrades and
                Josep Llados and 
                Cristina Canero},
  title     = {Recurrent Comparator with attention models to detect counterfeit documents},
  booktitle={ICDAR}, 
  year      = {2019},
  journal={Proc. IEEE}
}

Authors

  • Albert Berenguel (@aberenguel) Webpage

Acknowledgements

The initial idea of this paper was initially from Attentive Recurrent Comparators paper. Special thanks to Pranav Shyam and Sanyam Agarwal for their Theano and Pytorch implementation in which this work is based on. Additionally the ideas from Deep Co-attention based Comparators For Relative Representation Learning in Person Re-identification paper were adapted. For more details: https://github.com/pranv/ARC https://github.com/sanyam5/arc-pytorch

Included relevant citations

@article{DBLP:journals/corr/ShyamGD17,
  author    = {Pranav Shyam and
               Shubham Gupta and
               Ambedkar Dukkipati},
  title     = {Attentive Recurrent Comparators},
  journal   = {CoRR},
  volume    = {abs/1703.00767},
  year      = {2017},
  url       = {http://arxiv.org/abs/1703.00767},
  timestamp = {Wed, 07 Jun 2017 14:42:50 +0200},
  biburl    = {http://dblp.uni-trier.de/rec/bib/journals/corr/ShyamGD17},
  bibsource = {dblp computer science bibliography, http://dblp.org}
}
@article{wu2018deep,
  title={Deep Co-attention based Comparators For Relative Representation Learning in Person Re-identification},
  author={Wu, Lin and Wang, Yang and Gao, Junbin and Tao, Dacheng},
  journal={arXiv preprint arXiv:1804.11027},
  year={2018}
}