In this article you will find out more about xAI and how it is used to detect type of lung disease using x-ray images.
What is xAI?
This project started as a freelancing job. Our customer was a student of computer science in the US who got an assignment to create an AI that will be able to differentiate types of lung diseases using x-ray images. This convolutional neural network can differentiate between normal lungs, viral infections, bacterial infections, and smokers’ lungs.
In the development of CNN, we used Tensorflow. Since we did not have the same number of images for each type of disease, we used a data analysis technique called data augmentation to fill out the “missing” data and to ensure that the AI would have the same number of images to learn from. This is important because if a particular subset has a lower quantity of images, the network would be biased to a subset that has a higher quantity.
The model was trained using more than 5000 different images and has approximately 78% accuracy.
An example of input given to the neural network…
…and the output received.
You’ve read it all
Thanks for reading the article. If you have any questions, feel free to contact us.