Artificial Intelligence in medical image computing

Artificial Intelligence (AI) has a history of over 60 years, and deep learning represents the state-of-the-art AI. Deep neural network, which is the core of current deep learning models, is inspired by the human visual system. It’s therefore no big surprise to see that deep learning is making great breakthroughs in analyzing imaging data, as demonstrated by the many important applications, such as self-driving cars, face recognition, pedestrian detection. AI is showing transformative impact across many industries. Medical research and healthcare is no exception, and perhaps even more so, indeed healthcare systems have already started to move to the digital world since more than 20 years ago. Modern imaging technologies that allow us to see ourselves in unprecedented details. Physicians are increasingly relying on imaging data for diagnostic tasks. AI methods have emerged as valuable tools for medical image analysis in image detection, segmentation, registration, enhancement, synthesis, diagnosis, prognosis, etc. The research into AI-driven imaging diagnostics has great potential to improve the quality of clinical care, and as a consequence save lives and reduce healthcare costs and risks. A clear understanding of the strengths, limitations and applications of AI methods in imaging diagnostics will help us adapt to changes in the time of AI.