Research Article
COVID-19 Semantic Pneumonia Segmentation and Classification Using Artificial Intelligence
Algorithm 1
Preprocessing Configurations.
(i) | Input: Standard CXR Images (Images) | (ii) | Output: Processed CXR Images (OutImg) and Masks (msk) | (1) | Start Procedure | (2) | for im = 1: length (Images) | (3) | img = readIMG (Images [im]) | (4) | img = isGray (img) | (5) | img = reshape (256, 256) | (6) | img = adapthisteq (img) | (7) | End For | (8) | imgSize = [256, 256] | (9) | ncls = 2//number of classes | (10) | net = resnet50//pre-trained network | (11) | seg = deeplabv3plusLayers (imgSize, ncls,net) | (12) | opts.trainOptions (“softmax,” ‘MiniBatchSize,” 10, “MaxEpochs”, 30) | (13) | net = trainNetwork (Images, seg, opts) | (14) | msk = semanticSegmentation (Images, net) | (15) | return img, msk | (16) | End Procedure |
|