jump to navigation

AI Detects Rare Syndromes From Images January 9, 2019

Posted by stuffilikenet in Applications, Photography, Science, Star Trek Technology.
trackback

Rare disorders often present with patterned discoloration of the epidermis, distortion of features and other visibly-detectable aberrations in appearance. Marfan’s syndrome presents with long, flexible body type. Noonan syndrome may present wide-set eyes, and Down’s syndrome is well-known to nearly everyone. Now, researchers have developed a facial analysis framework, DeepGestalt, using computer vision and deep learning algorithms, that quantifies similarities to hundreds of genetic syndromes based on unconstrained 2D images. DeepGestalt is currently trained with over 26,000 patient cases from a rapidly growing phenotype-genotype database, consisting of tens of thousands of validated clinical cases, curated through a community-driven platform. DeepGestalt currently achieves 91% top-10-accuracy in identifying over 215 different genetic syndromes and has outperformed clinical experts in three separate experiments.

In results published in Nature Medicine, DeepGestalt  outperformed doctors in diagnosing patients with Angelman syndrome and Cornelia de Lange syndrome versus other disorders, and in separating patients with different genetic subtypes of Noonan syndrome.

It’s a neat study in that it controls for a bunch of conditions including ethnicity and gender, so it’s a bit more robust than previous studies.

 

Homework: [PDF] https://arxiv.org/abs/1801.07637

Comments»

No comments yet — be the first.

Leave a Reply

Fill in your details below or click an icon to log in:

WordPress.com Logo

You are commenting using your WordPress.com account. Log Out /  Change )

Google+ photo

You are commenting using your Google+ account. Log Out /  Change )

Twitter picture

You are commenting using your Twitter account. Log Out /  Change )

Facebook photo

You are commenting using your Facebook account. Log Out /  Change )

Connecting to %s

%d bloggers like this: