Selected article for: "binary classification task and classification task"
Author: Xuehai He; Xingyi Yang; Shanghang Zhang; Jinyu Zhao; Yichen Zhang; Eric Xing; Pengtao Xie
Title: Sample-Efficient Deep Learning for COVID-19 Diagnosis Based on CT Scans
  • Document date: 2020_4_17
    • ID: l3f469ht_32
    Snippet: Considering that our dataset is small and our task is binary classification, to study whether there is indeed overparametrization in the traditional ImageNet models when applied to COVID-19 diagnosis, in addition to these largesized architectures, we also design a light-weight architecture as shown in Figure 3 . The basic building block for this network is the combination of a 2d-convolution with an ReLU activation. The block repeats four times. .....
    Document: Considering that our dataset is small and our task is binary classification, to study whether there is indeed overparametrization in the traditional ImageNet models when applied to COVID-19 diagnosis, in addition to these largesized architectures, we also design a light-weight architecture as shown in Figure 3 . The basic building block for this network is the combination of a 2d-convolution with an ReLU activation. The block repeats four times. There is an additional batch normalization layer in the first block and an average pooling layer after the third block. We refer to this network as CRNet, whose structure is as follows: (conv32-bn-relu), maxpool, (conv64-relu), maxpool, (conv128-relu), maxpool, global avgpool, and classification layer. The conv(n) represents a 2d convolutional layer with n output channels with a kernel size of 7×7 and a stride of 1. The bn denotes a 2d batch normalization [50] layer. The relu stands for an ReLU layer. The maxpool stands for max pooling with a kernel size of 3×3 and a stride of 2. The global avgpool is an average pooling layer with a kernel size of 2×2 and a stride of 2.

    Search related documents:
    Co phrase search for related documents
    • average block pooling layer and classification layer: 1
    • average block pooling layer and pooling layer: 1
    • batch normalization and binary classification: 1
    • batch normalization and block batch normalization layer: 1
    • batch normalization and block pooling layer: 1
    • batch normalization and classification layer: 1
    • batch normalization and convolutional layer: 1, 2
    • batch normalization and pooling layer: 1, 2, 3, 4, 5
    • batch normalization and relu activation: 1, 2
    • batch normalization and relu layer: 1, 2, 3
    • batch normalization layer and block batch normalization layer: 1
    • batch normalization layer and classification layer: 1
    • batch normalization layer and convolutional layer: 1, 2
    • batch normalization layer and pooling layer: 1, 2
    • batch normalization layer and relu activation: 1, 2
    • batch normalization layer and relu layer: 1, 2
    • binary classification and classification layer: 1, 2
    • binary classification and convolutional layer: 1, 2, 3, 4
    • binary classification and ImageNet model: 1