Computer Vision News - January 2022

68 MedTech Spotlight News Medical Imaging News has found great new stories, written somewhere else by somebody else. We share them with you, adding a short comment. Enjoy! AI Accurately Predicts Who Will Develop Dementia in Two Years ThisAIpredictingwhowill developdementia in two years with 92%accuracy is important stuff, because it reduces the guesswork in clinical practice and significantly improve the diagnostic pathway, helping families access the support they need as swiftly and as accurately as possible. It reduces by 80% the number of false positives reducing the unnecessary distress that a wrong diagnosis could cause, by spotting hidden patterns in the data and learning who is most at risk. This research from the University of Exeter used data from more than 15,300 patients in the United States. Read More Mind-controlled Robots Now One Step Closer Two groups of researchers at EPFL Lausanne havedevelopedanML program that commands a robot arm obeying to a human via electrical signals from the brain . ThisML uses inverse reinforcement learning to learn commands coming a tetraplegic patient wearing a headcap equipped with electrodes scanning the patient’s brain activity. When the robot makes an incorrect move, the patient’s brain emits an error message through a clearly identifiable signal, which indicates to the robot “ No, not like that! ” The robot will try again with another course until it learns the patient’s intention. Watch the video! AI-Driven Medical Imaging May Help Fight Against Rectal Cancer A Case Western Reserve University -led team is figuring how AI can help rectal cancer patients . Apparently, many of them cannot rely on chemotherapy or radiation, so most patients have to undergo invasive surgery. Researchers want to avoid patients being overtreated , so their AI will tell clinicians right up front - based on a routine MRI scan - if a patient will do well with only chemoradiation, without having this serious surgery . They will test their radiomics (an AI that extracts many features from medical imaging using algorithms on large data) on about 450-500 patients. Read More