MEDICAL IMAGING VISUALIZATION & DIAGNOSTICS WITH AI

Ruth Yakubu
Microsoft Azure
Published in
3 min readFeb 19, 2021

Startups transforming how medical images are used in Health AI

Radiology imaging

Developing high accuracy machine learning models is crucial in helping medical professionals to diagnose diseases or health conditions. To successfully identify health conditions, hospitals need medical images generated from X-Rays, Ultrasounds, MRI, CT Scans, or Mammographs. However, many of these conditions are difficult to identify with a human eye or in a timely manner. The advancements in deep learning algorithms enable AI to detect abnormalities in these medical images. To train the AI model, a large amount of time is spent in curating the medical images to be used in training the model. One of the major challenges of building highly reliable and accurate AI models requires curating massive data set of images for machine learning. To find and fix any issues with the data images, data visualization is needed. In most cases, having poor machine learning results is linked to poor training data. In a healthcare setting, performing due diligence of cleansing the data is needed in making sure the quality of the training data is good in order to save lives.

One of the companies leading the breakthrough of medical image curation and visualization for higher accuracy in diagnostics in the Health industry is Zegami. Zegami’s AI platform is allowing scientists or clinicians to look for trends and outliers in thousands of medical images.

We are excited to be hosting Zegami on our next episode of the Tech Exception show, where we have conversations with startups disrupting the AI space to learn how they are pushing the boundaries of innovation.

Join us LIVE on February 24th from 2pm — 2:30pm PT on the Tech Exceptions show where we’ll be having a conversation with Roger Noble, CEO of Zegami, backed by Oxford Science & Innovation for their AI enabled image analysis platform for scientific & medical imaging. Their solution makes it easy to view, manage, graph, categorize, analyze, and spot outliers in heaps of imaging data quickly and intuitively. In this session, we will discuss techniques the company uses to drill into thousands of images to find detailed patterns and insights using dynamic views, filtering, and graphing to find potential biases or erroneous entries for computer-vision based AI training.

Be part of the live conversation of the Tech Exceptions show on Microsoft Learn TV. This will be an interactive show, so we invite you to post questions, learn, engage, and connect with the startup leaders shaping the future of technology.

Tech Exceptions show

Watch Live on Learn TV February 24th 2:00–2:30pm PT!

Mark your calendars! Make sure to follow Ruth Yakubu and Tech Exceptions on social media for more information and updates! Subscribe to the Tech Exceptions Show for lots more exciting episodes coming up!

--

--