jacklxc / MUDE

Learning Multi-level Dependencies for Robust Word Recognition

Learning Multi-level Dependencies for Robust Word Recognition

Pytorch implementation of MUDE. Some parts of the code are adapted from the implementation of scRNN.

For more details about MUDE, Please read our paper. If you find this work useful and use it on your research, please cite our paper.

@article{wang2019learning,
  title={Learning Multi-level Dependencies for Robust Word Recognition},
  author={Wang, Zhiwei and Liu, Hui and Tang, Jiliang and Yang, Songfan and Huang, Gale Yan and Liu, Zitao},
  journal={arXiv preprint arXiv:1911.09789},
  year={2019}
}

Usage

Our repository is arranged as follows:

    data/ # contains the original datasets
    experiment.py #run experiment to produce Table 2 and 3
    model.py # comtains the MUDE
    utils.py # utility functions
    generalization_pipeline.py # produce Table 4 and 5

When data is ready, the code can directly run with PyTorch 1.0.0.