Machine Learning Use Case in Dentistry

We always talk about different techniques of machine learning and how they can be implemented. But today, I want to share with you a real world use case of how we used ML as part of a project destined to help dentists.

When dentists are going to do an implant for a patient, they first do some analysis on a radiograph of the exact area where they plan to do the implant.

The analysis that they do is as follows:

• They look at the bone area of the mandible (lower jaw) and they look at the area of the canal.
• They then define some metrics:
  • Height: a segment that goes from the upper side of the canal all the way to the upper side of the bone.
  • Width: a segment that's perpendicular on the height segment at a point that's defined by a distance in millimeters from the upper side of the height segment.
• Finally, they use the length of the height segment and the length of the width segment to categorize the case into one of 25 classes.

Based on the metrics above, they define how the dental implant surgery should be.

The only part that requires machine learning in this use case, is the radiograph analysis to highlight (segment) the bone area and the canal area.

This is a real world use case of an application that uses ML and that can provide value to the end users, who are dentists in this case.

In fact, we've built such an app for one of our clients. You can see below a screenshot of the final desktop app that works without a need to access the internet:

Are you looking to build an app (desktop or mobile or web app) for medical imaging? Feel free to reach out to us at: contact@pycad.co

​

Deep Learning for 3D Medical Data

Are you interested in using deep learning with 3D MRI data?
​
Then check out my article where I share 6 different deep learning models that you can use with 3D MRI data.
​
The article contains links to papers and code for each model.

​

Tweet of the Day

​

Bennett Landman (Vanderbilt MASI Lab) @MASI_Lab We spend a lot of time making algorithms work in 3-D. Does 2-D still matter? Especially for CT, why would anyone use single slice imaging? 1 word: Dose. In new work published in @SPIEtweets Journal of Medical Imaging, Qi Yang shows that we use domain adaptation to effectively. /1 3:49 AM • Sep 20, 2023 3 Retweets 4 Likes Read 1 replies

​

​

Meme of the Day 😂