Example Application
To demonstrate the PowerAI Insights streamline processes on how to 1) label image, 2) train, and 3) deploy the model, a classification of COVID-19 in the chest X-ray image is given here. With the high availability of large-scale annotated X-ray image datasets, great success has been achieved using convolutional neural networks (CNN) for medical diagnosis. \cite{gaber2020} \cite{hassanien2020} Yet the applied models in these previous studies involve some advanced algorithms, such as transfer learning from other generic object recognition tasks, which makes them less intuitive for subject matter experts with limited deep learning skills. In the following, we will show how PowerAI Insights help train models only with a few clicks.
Dataset Characteristics
The dataset is obtained from a Github repository publicly released by Skytells.\cite{skytells-research} Figure 1 shows a sample of images with 4 categories, the training dataset contains 880 normal, 60 COVID-19, 650 Bacteria Pneumonia, and 412 Viral Pneumonia images. All images were of the same size and stored in JPEG format with 512*512 pixels.