Next-generation AI for frictionless visual telemedicine

Produce accurate skin assessments on patients using a detailed 5-step image analysis AI.

hero section

About Us

developed by a team of engineers and doctors from MIT, Harvard, and ASU

Luis Soenksin

Luis Soenksin PhD

Serial entrepreneur and medical device expert. MIT and Johns Hopkins BME graduate currently acting as MIT’s first Venture Builder in Artificial Intelligence and Healthcare.

Omar Badri

Omar Badri M.D

Serial-entrepreneur, Internal Medicine, Co-founder of Medumo - company to guide, track, triage and educate patients (acquired by Phillips 2019). Practices at Brigham and Women’s Hospital (BWH) and Harvard Medical School.

Alex Alimov

Alex Alimov

Full-stack Software developer +10Y Experience in Mobile IOS, Android, Cloud.

Product

5-step image analysis for accurate skin assessments

01

Image quality assessment

Every image is assigned a quality score to determine if an accurate analysis can be performed by automated algorithms

Step 1
Step 2

02

Body Location mapping

Visible body parts in submitted images are automatically recognized, mapped and tracked

03

Fitzpatrick skin type

Upon face detection, a skin color palette is recognized to calculate the average phototype of the patient. Other features such as age, sex and ethnic skin are automatically infered to facilitate reporting

Step 3
Step 4

04

Image-based Differential Diagnosis

Our pre-trained deep learning algorithm receives image crops from received photos and produces arank ordered by diagnostic probability across 250 diseases with visual presentations

05

Question-based Refinement

A short list of highly focused questions is selected based on the determined differential diagnosis. These questions are designed to improve diagnosis probability and work through the short list of possible diagnoses

Step 5
Step 6

06

Final Report

Physician is presented with full report of aggregated information to allow for best decision making process in least amount of review time