2013

Sparse arrays of signatures for online character recognition

Graham, Benjamin

Understand

In mathematics the signature of a path is a collection of iterated integrals, commonly used for solving differential equations.

  • We show that the path signature, used as a set of features for consumption by a convolutional neural network (CNN), improves the accuracy of online character recognition---that is the task of reading characters represented as a collection of paths.
  • Using datasets of letters, numbers, Assamese and Chinese characters, we show that the first, second, and even the third iterated integrals contain useful information for consumption by a CNN.
  • On the CASIA-OLHWDB1.1 3755 Chinese character dataset, our approach gave a test error of 3.58%, compared with 5.61% for a traditional CNN [Ciresan et al.].

Built on

  • Integration of paths—a faithful representation of paths by non-commutative formal power series

    Kuo-Tsai Chen · 1958

    Earlier work this paper cites.

  • System control and rough paths

    Terry Lyons and Zhongmin Qian · 2002

    Earlier work this paper cites.

  • Learning methods for generic object recognition with invariance to pose and lighting

    Yann LeCun, Fu-Jie Huang, and Leon Bottou · 2004

    Earlier work this paper cites.

  • A study on the use of 8-directional features for online handwritten Chinese character recognition

    Z. L. Bai and Q. A. Huo · 2005

    Earlier work this paper cites.

Similar

  • Sound compression—a rough path approach

    T. Lyons and N. Sidorova · 2005

    Cited alongside, same era.

  • Efficient learning of sparse representations with an energy-based model

    Original

    Marc’Aurelio Ranzato, Christopher S. Poultney, Sumit Chopra, and Yann LeCun · 2006

    Cited alongside, same era.

  • Differential equations driven by rough paths

    Terry J. Lyons, Michael Caruana, and Thierry Lévy · 2007

    Cited alongside, same era.

  • Uniqueness for the signature of a path of bounded variation and the reduced path group

    Ben Hambly and Terry Lyons · 2010

    Cited alongside, same era.

  • The MNIST database of handwritten digits

    Yann Lecun and Corinna Cortes

    Cited in the paper.

Then

  • CASIA online and offline Chinese handwriting databases

    C.-L. Liu, F. Yin, D.-H. Wang, and Q.-F. Wang · 2011

    Later among the works it cites.

  • Multi-column deep neural networks for image classification

    D. Ciresan, U. Meier, and J. Schmidhuber · 2012

    Later among the works it cites.

  • Improving neural networks by preventing co-adaptation of feature detectors

    Original

    Geoffrey E. Hinton, Nitish Srivastava, Alex Krizhevsky, Ilya Sutskever, and Ruslan Salakhutdinov · 2012

    Later among the works it cites.

  • UCI machine learning repository, 2013

    K. Bache and M. Lichman · 2013

    Closest in time.

Beyond the bibliography

alphaXiv searches the wider corpus for related work and actual follow-ups.

Open on alphaXiv

alphaXiv is searching for related work…