Selected article for: "high number and low number"

Author: Saban Ozturk; Umut Ozkaya; Mucahid Barstugan
Title: Classification of Coronavirus Images using Shrunken Features
  • Document date: 2020_4_6
  • ID: 2l1zw19o_43
    Snippet: is the (which was not peer-reviewed) The copyright holder for this preprint . https://doi.org/10.1101/2020.04.03.20048868 doi: medRxiv preprint For this reason, the feature extraction method is most suitable for hand-crafted methods. However, the fact that the number of images in the dataset is quite low and the number of sample differences between classes is very high (almost 90% belong to only one class) will create a problem for classifier alg.....
    Document: is the (which was not peer-reviewed) The copyright holder for this preprint . https://doi.org/10.1101/2020.04.03.20048868 doi: medRxiv preprint For this reason, the feature extraction method is most suitable for hand-crafted methods. However, the fact that the number of images in the dataset is quite low and the number of sample differences between classes is very high (almost 90% belong to only one class) will create a problem for classifier algorithms.

    Search related documents:
    Co phrase search for related documents
    • classifier algorithm and feature extraction method: 1
    • dataset image and feature extraction: 1, 2, 3, 4, 5, 6, 7
    • feature extraction and sample difference: 1