Privacy Preserving Abnormality Detection: A Deep Learning Approach

The development of AI has made some major advances in recent years and its potential appears to be promising. In the healthcare sector, scientific competitions like ImageNet Large Scale Visual Recognition Challenges are providing evidence that computers can achieve human-like competence in image recognition. There are numerous computer models in medical diagnosis to help physicians. Among different models, deep learning algorithms, in particular convolutional neural networks are among the first choices for medical images analysis. This paper use one of the largest dataset of open-source musculoskeletal radiographs (MURA) for abnormality detection of thousands of musculoskeletal radiographs based on the deep learning to build models for detecting and localizing the abnormalities.