Deep Learning Algorithm for COVID-19 Classification Using Chest X-Ray Images

Algorithm 1

Proposed method.

1. Input data: for CXR real images, let be input image and be output label. = {normal, pneumonia, and COVID-19}.

2. Output data: for output label y, the result may return as normal, pneumonia, or COVID-19.

3. Preprocessing phase: CXR images are modified to the height and width dimension of .

4. Training phase: for the number of training iterations for steps:

  (i) Sample minibatch of noise sample {,…., } from noise prior )

  (ii) Sample minibatch of example {,….,⁾} from data generation distribution

  (iii) The real image is transferred to the discriminator. The discriminator is updated by ascending its stochastic gradient using the transfer model

end for

  (i) Minibatch of m noise sample {,…., } is sampled from noise prior

  (ii) The discriminator is updated by descending its stochastic gradient

end for

5. Testing phase: output label is generated.