Computer vision technology continues to expand its use cases in healthcare and medicine. In this post, we will touch on some exciting example use cases for vision in healthcare and medicine and provide some resources on getting started applying vision to these problems.

What Makes a Problem Solvable with Computer Vision?

Problems that are solvable in healthcare and medicine share similar characteristics:

1) The solution involves automatically transforming images or video into actionable information.

2) The solution operates probabilistically. Computer vision models are trained to make predictions that are estimates in nature - this means that problems that can be handled with some error rate tend to be a good fit.

3) The image or video input has limited variation. Problems that have a greater variation in scene are generally harder for a computer to learn how to solve.

Of course, the challenges of every problem in computer vision will be unique, but they will likely fall into one of these three categories.

Using Computer Vision to QA Medical Equipment and Devices

Computer vision is useful to assure quality assurance. Medical devices and equipment need to be manufactured within specifications and manual monitoring their quality can be tedious and expensive.

Before administering a treatment via syringe, we may a computer to automatically detect whether the quantity of liquid is within an acceptable error range. We may also want an algorithm to detect whether all pieces of the syringe are intact and accounted for.

Ensuring the quality assurance of syringe treatments (source)

The quality assurance process naturally lends itself to computer vision because pieces of equipment are manufactured in similar fashion, reducing the variability that your computer vision model will need to learn.

Using Computer Vision to Triage and Diagnose Ailments

X-Ray - Computer vision for reading MRI and X-Rays has become a token example of computer vision in healthcare. Machine learning technology now enables doctors to screen and reason about these medical images more effectively because the computer can pick up on patterns that a human may not.

Identifying areas of malady in X-Ray imagery (source)

Dental - Computer vision can also be used to augment triage and diagnosis in dental care. Models are trained to recognize malformations in teeth and gum tissue and automatically make screening recommendation.  

Skin - Images of a patient's skin are automatically scanned with computer vision technology to screen for ailments such as skin cancer.

Monitoring Physical Therapy Exercise

Computer vision is used to automatically monitor the behavior of patients in physical therapy exercise. Patients install an application that takes a video of them performing their exercise. The amount of sets and repetitions are automatically identified by the computer vision algorithm, motivating the patient and providing telemetry to the care provider.

Using Computer Vision to Accelerate Medical Research and Production

Computer vision is used to augment medical research procedures. For example, vision algorithms are trained to identify and count different cell types at the microscopic level.

Detecting cells automatically

Vision techniques are used to automatically sort particles in fluids and count pills.

Automatically counting pills with computer vision

Getting Started Using Computer Vision in Healthcare and Medicine

These are but a few examples of how computer vision is being applied to the healthcare and medicine today.

To get started on applying computer vision to your task, we recommend reviewing the section above about "What makes a problem solvable with computer vision?"

If you think your task reasonably fits into those categories, the next step is to begin sourcing training images and then to train and deploy a model.

Here are some great resources for next steps:

And last, but never least:

As always, happy training!