Convolutional Neural Networks and Machine Learning/Deep Learning in general, has been implemented in a variety of fields. Medical imaging is one of the most intriguing ones, and so I decided to use AI to approach the problem of cell segmentation in microscopic images.

The What

Consider the image below — a microscope slide that contains a bunch of cells, including one white blood cell. If a person wanted to analyze the white blood cell, they would first have to figure out which one the white blood cell, actually is. You would have to do the same thing if you were to have a computer perform analysis on the white blood cell. The problem of separating parts of an image is called **segmentation. **There are a variety of ways to do this, but I performed semantic segmentation — assigning each pixel in the image to a class. In this case, the classes were white blood cell and not white blood cell. To gain a more thorough understanding of semantic segmentation is, I’d recommend checking out this great article.

#machine-learning #deep-learning #artificial-intelligence #tensorflow #segmentation