A platform for medical professionals streamlines uploading datasets using AI to entrain image recognition diagnostics.
In medicine, the speed and accuracy of diagnoses can make the difference in patient survival. Could algorithms help doctors achieve diagnoses faster? Indeed, we’ve already seen an example of machine learning being used to detect colorectal cancers.
MedicMind, from oDocs, has taken this approach, developing a platform that enables medical users to develop image recognition systems. This system is usable for a vast range of image-based diagnostics. The software uses Google’s open source Tensorflow software in its construction. Medical researchers will already have vast datasets of images for an array of diseases. MedicMind uploads the images and gives them scores by experts, thereby developing a diagnostic scale. This entrains an AI algorithm to ‘know’ when an image it sees is likely to be a diseased or healthy image at a much faster and more accurate rate than humans can achieve. As users add more images, MedicMind will only grow more accurate.
While we’ve covered examples of image recognition software before, it’s MedicMind’s offering of this Software-as-a-Service (SaaS) that will enable rapid uptake by a wide range of medical-based professionals. The MedicMind platform is simple to use. Users upload datasets through drag-and-drop menus and analytical tools are all already in place, so that researchers won’t need to code their own algorithms. The MedicMind platform stores data so that it is easily retrievable at any time. With published research already showing the success of machine learning image recognition diagnostics for diseases such as skin cancer, diabetic retinopathy, pneumonia, breast cancer and others, AI algorithms could soon become the norm across a range of medical disciplines.
Image recognition software is becoming increasingly prevalent. We’ve seen the technology being used in surveillance, real estate and to track responses to adverts. What other industries will benefit from using image recognition software?