Volume 10 Issue 2 - April 2018

  • 1. Birds species classification

    Authors : Potluri Rohit Sai Chowdary, L. Hemanth Reddy, K. Sugamya

    Pages : 347-352

    DOI : http://dx.doi.org/10.21172/1.102.56

    Keywords : KNN, SVM, LDA.

    Abstract :

    Birds- a body of endothermic vertebrates and classified by wings, beaks, belly patterns, tails, forehead, etc., Birds are found globally ranging from 2in to 9ft in size . There are about ten thousand living species of birds and are ranked as the world’s most numerically successful class of tetrapods. It’s almost impossible even for an ornithologist to classify each and every bird by looking at its features. We are making use of Caltech-UCSD Birds-200-2011 (CUB- 200-2011) for training and testing. We have used different algorithms for classifying different bird species into different classes and we have achieved the accuracy of 44.44% using LDA as compared to 43.7% using SVM, 33.58% using random forests, 20.4% using KNN, 19.22% using Naive Bayes.

    Citing this Journal Article :

    Potluri Rohit Sai Chowdary, L. Hemanth Reddy, K. Sugamya, "Birds species classification", https://www.ijltet.org/journal_details.php?id=930&j_id=4516, Volume 10 Issue 2 - April 2018, 347-352, #ijltetorg