The task of recognizing characters and numbers from handwritten images has a wide variety of applications in today's world In this paper, a deep hybrid learning approach using transfer learning to classify numbers from handwritten images of regional language Odia numbers was implemented. The performance across various optimizers on VGG16, and even among various traditional machine learning classifiers, were compared. Each of the networks was trained for 50 epochs, with ReduceLR on plateau as a callback mechanism to make the model learn the parameters better. Augmentation techniques such as image shearing and rescaling to further generalize the model.


    Access

    Check access

    Check availability in my library

    Order at Subito €


    Export, share and cite



    Title :

    Odia Handwritten Numeral Recognition: A Hybrid Modelling Approach


    Contributors:


    Publication date :

    2021-12-02


    Size :

    1062385 byte




    Type of media :

    Conference paper


    Type of material :

    Electronic Resource


    Language :

    English



    Script Independent Handwritten Numeral Recognition

    Kunte, R. S. / Sudhaker, S. R. D. / Visual Information Engineering Network (Institution of Engineering and Technology) | British Library Conference Proceedings | 2006


    A hybrid multiple classifier system of unconstrained handwritten numeral recognition

    He, C. L. / Suen, C. Y. | British Library Online Contents | 2007


    A Hybrid Multiple Classifier System of Unconstrained Handwritten Numeral Recognition

    He, C. L. / Suen, C. Y. | British Library Online Contents | 2005


    Approach to subclass division and discrimination for handwritten numeral recognition [4554-48]

    Zhu, X. / Shi, Q. / International Society for Optical Engineering | British Library Conference Proceedings | 2001


    A New Set of Moment Invariants for Handwritten Numeral Recognition

    Pan, F. / Keane, M. / IEEE; Signal Processing Society | British Library Conference Proceedings | 1994