Pearse Keane working with AI on imaging

This piece describes more of what might be done than what is done in automated medical imaging analysis by artificial intelligence.  “As researchers and clinicians have excellent access to images and patient data within their own institutions, they might be able to design automated machine learning models for internal research, triage, and customised care delivery. Automating these processes might avert the need for costly external prospective validation across imaging devices, patient populations, and clinician practice patterns. By contrast, large-scale, standardised, deep learning algorithms will necessitate worldwide, multivariable validations of expertly tuned models. Thus, there is considerable value to these small data approaches customised to a specific geographical patient population that a given clinic might encounter. This might be where automated deep learning finds its niche in the medical field.”    MORE

Image Credit: Moorfields Eye Hospital