Front Inner Page - Volume 4 No.3 June 2017

  • » Back to Index

  • Title

    :

    A Feature Learning and Object Recognition Framework for Underwater Fish Images with Segmentation

    Authors

    :

    O. Letcy Fernando 1, J. Rexy2

    Keywords

    :

    Fish recognition framework, Fishery survey applications, Error resilient classifier, Binary class hierarchy

    Issue Date

    :

    June 2017

    Abstract

    :

    Live fish recognition is one of the most decisive elements of fisheries analysis where the massive amount of data is promptly assimilated .Diverse from wide-ranging scenarios, experiments to underwater image recognition are forwarded by poor image quality, uncontrolled objects and environment, and efforts in obtaining demonstrative illustrations. In addition, most existing feature extraction techniques are caught up from automation due to human observation. Hence, we propose an underwater fish recognition framework that comprises of an entirely unsupervised feature learning procedure and an error-resilient classifier. Object parts are modified based on saliency and slackening cataloging to match objective parts appropriately.

    Page(s)

    :

    1-7

    ISSN

    :

    2347- 4734

    Source

    :

    Vol. 4, No.3, June 2017

    Download

    :


  • » Back index