Selected article for: "adam optimizer and logarithmic softmax layer"

Author: Rahul Kumar; Ridhi Arora; Vipul Bansal; Vinodh J Sahayasheela; Himanshu Buckchash; Javed Imran; Narayanan Narayanan; Ganesh N Pandian; Balasubramanian Raman
Title: Accurate Prediction of COVID-19 using Chest X-Ray Images through Deep Feature Learning model with SMOTE and Machine Learning Classifiers
  • Document date: 2020_4_17
  • ID: 59ghorzf_20
    Snippet: • A ResNet152 22 model was trained for the classification of Pneumonia and Normal patients. ResNet is known to be a better deep learning architecture as it is relatively easy to optimize and can attain higher accuracy. Due to a large number of layers in the network architecture, it has high time complexity. This complexity can be reduced by utilizing a bottleneck design. Further, there is always a problem of vanishing gradient, which is resolve.....
    Document: • A ResNet152 22 model was trained for the classification of Pneumonia and Normal patients. ResNet is known to be a better deep learning architecture as it is relatively easy to optimize and can attain higher accuracy. Due to a large number of layers in the network architecture, it has high time complexity. This complexity can be reduced by utilizing a bottleneck design. Further, there is always a problem of vanishing gradient, which is resolved using the skip connections in the network. Finally, the last fully connected (FC) layer of the network is followed by a Logarithmic Softmax layer with Adam optimizer to optimize the neural network.

    Search related documents:
    Co phrase search for related documents