Once the images are downloaded, the first step is to import them into the platform and assign proper categories to each image. The images can be imported one category at a time and assigned labels (such as covid19, bacterial, viral, and normal in our case).  The interface for importing and categorizing images is shown in Figure \ref{157103}. Images with format JPEG, PNG, and DICOM are supported. The models used by IBM Visual Insights have limitations on the size and resolution of images (1-2 megapixels). High-resolution images need to be divided into smaller sub-images if they require fine details, or down-sampled if they have coherent structural information that can not be divided. During the training, the images will be automatically scaled to the appropriate dimensions for the model to use. For example, the 400x300 pixels images will be scaled to meet the GoogLeNet requirement of 224x224 pixels.