We are working on a cloud based platform called Medicmind (http://www.medicmind.tech) that allows medical researchers to train their own AI without needing to code. For example you can drag and drop 1000 images of glaucoma and non glaucoma images into medicmind and click train and it’ll train a neural network using Googles state of the art Inception V3. You can then deploy the trained neural network in numerous ways like on a smartphone (http://medicmole.com) or via our Medicmind API (https://www.doc2us.com/ ) or just drag an drop directly into Medicmind.
This is really great stuff! Great opportunities here in the radiology space.
opengraphobject:[351561250283754 : https://www.doc2us.com/ : title=“Doc2Us” : description=“”]
If you’ve got your own radiology dataset you can easily train on Medicmind if you want to try it out. Medicmind is in beta and we’re keen to get users trying it out.
opengraphobject:[351561250283754 : https://www.doc2us.com/ : title=“Doc2Us” : description=“”]
I’ve sent a link and your details to the radiology manager at the DHB I am on secondment at.
opengraphobject:[351561250283754 : https://www.doc2us.com/ : title=“Doc2Us” : description=“”]
Excellent. If you need any help dont hesitate to ask. There’s a brief video of creating a grading classifier at https://www.youtube.com/watch?v=ukT671xClrE Grading classifiers are basically giving a number between 0 and 1 with a probability of a pathology like glaucoma or not. There are videos for all the types of Medicmind classifiers like multiclass (whether an image is a dog, cat, bird)
opengraphobject:[351561250283754 : https://www.doc2us.com/ : title=“Doc2Us” : description=“”], opengraphobject:[351561250386064 : https://www.youtube.com/watch?v=ukT671xClrE : title=“Part 2 MedicMind Using the Graded Classifier” : description=“MedicMind offers a cloud based medical AI development platform. The video offers instruction and examples of its application. Visit http://www.medicmind.tech/”]