There is a dire shortage of radiologists globally. In the US, there is 1 radiologist per 10,000 people, whereas the ratio is much lower in many developing countries. Moreover, X-rays and patient electronic medical records (EMR) are stored in different systems, which has led to the current disjoint workflow for radiologists to diagnose patient X-rays.
ChestX.ai is focused on addressing the previously stated challenges to empower radiologists in their X-ray diagnosis process. Using more than 110,000 images from the NIH and newly aggregated COVID-19 chest X-ray datasets, ChestX.ai applies the DenseNet-121 Convolutional Neural Network architecture and Grad-CAM heat map visualizations on chest X-rays to offer AI findings to radiologists, as well as integrating patient EMR data to improve diagnosis efficiency. ChestX.ai intends to use our MVP to contribute to peer and industry learning about connecting machine learning with patient medical records to also empower other medical professionals.