Chest X-ray is the most common medical imaging method globally and is of particular importance in the screening, diagnosis, and treatment of many life-threatening diseases. Most chest X-ray interpretation AI algorithms are developed based on digital X-ray images, whereas most developing countries still use X-ray films. One solution for the large-scale deployment of these AI models is to leverage the ubiquity of smartphones to capture and automatically read X-ray films. However, the challenge is how to make the reading of X-ray images taken with a smartphone as good as the quality of digital X-ray images because X-ray images taken with photos usually have visual artifacts not commonly found in digital X-rays. In order to encourage the development of deep-learning AI models with high accuracy when reading X-ray images that are naturally noisy due to being captured from digital X-ray images or hard films by smartphones, VinBrain and Stanford University co-hosted the CheXphoto Contest (https://stanfordmlgroup.github.io/competitions/chesphoto/#test_and_validation). The CheXphoto competition measures the performance of AI models against the CheXpert X-ray image dataset. The contest uses a private dataset to formally evaluate the performance of the models. The teams upload the code that runs the model to Codalab, which will be run on an undisclosed test set. This process is intended to ensure the transparency and fairness of the test results.

According to Mr. Truong Quoc Hung - General Director of VinBrain, the cooperation with Stanford University to organize the CheXphoto contest is meant to promote the research, development, and extensive deployment of AI solutions to support medical imaging diagnostics in all environments, all circumstances. We are all living in a flat and open world. ChexPhoto is our way to contribute to the mutual sharing of information, data and knowledge, in order to accelerate the application of high technology into healthcare to improve people's lives and productivity.

Stanford University is one of the strategic partners of VinBrain since the early days of its establishment with the common goal of improving people's lives and productivity through the research and application of modern technologies.