Chest X-Ray Analysis Using Machine Learning for Detection of Abnormalities

Earlier SOLVVE wrote about one of the projects done by the machine learning department using object detection and recognition to automatically extract and register game results in eSports. This time we would like to show how similar techniques can be used to detect and identify different kinds of diseases on chest X-rays. Below you will also be able to try our tool for yourself.

How does it work?

In simple terms, object detection is a part of computer vision. It is a processing technology that can understand if an object or a person is present in a picture. For example, it is widely used in surveillance systems. Among other things,  facial recognition technology uses these techniques to find people. You can learn more about object detection in detail from our dedicated article here.

However, when we talk about X-ray pictures, the key tool of this detection mechanism is a deep convolutional network called Faster R-CNN. It’s primary goal is to accurately detect objects of interest on images. In our case it is all sorts of image distortions that can be associated with the presence of a certain decrease on a chest X-ray.

What are the potential applications of this model?

On the one hand, such tools are very useful in helping doctors process patient data faster by getting a preliminary diagnosis. However, it is worth noticing that at the current stage none of the ML-backed solutions can fully substitute a medical professional despite extremely high levels of accuracy of detection and identification. Still, automatic image processing can significantly boost doctors’ productivity and performance. 

On the other hand, this tool can also be adapted for other types of images. Or, machine learning engineers can also work on making the recognition quality of images better by increasing the data volume in training sets. 

I would like to have this technology for my business

SOLVVE machine learning department has created a tool to scan chest X-rays for abnormalities and try to identify a variety of diseases associated with image distortions. You can try it for yourself here. 

The machine learning department at SOLVVE already has hands-on experience with this instrument and can help you plan, develop and deploy computer vision and image recognition features for your product or application. If you have any questions or ideas about using machine learning in your business, do not hesitate to contact us. Let us make it happen!