CheXpert: Radiology Image Classification via Machine Learning

Stanford researchers have developed CheXpert which can reduce noise and identify several pathologies on x-rays with very high accuracy via machine learning. CheXpert can read photos of x-rays from a mobile phone and is robust to noise. The trained neural network is computationally light enough to run on mobile platforms such as cell phones or low-power computers. This can enable users anywhere, even in remote regions or without access to significant computing resources, to utilize the neural network for the radiology image classification. The dataset is open source and can be a standard benchmark to evaluate performance of chest radiograph interpretation models.

Stage of Development