Pixel Perfect

The utilization of image segmentation can be seen in a variety of industries – medical diagnosis, autonomous driving, and face recognition. Image pixels can be interpreted as nodes of a here graph with edges connecting neighboring pixels; this approach to viewing images allows us to use different network optimization methods to conduct image analysis. In this project, I utilize these methods to do image segmentation to split an image into the foreground and background. This can be completed several ways; I include 2 in my project. The first is naïve thresholding, where the parameter choices are set somewhat arbitrarily. The other is using the min-cut/max-flow approach, where we compute matrices to differentiate between foreground and background pixels, as well as the weights on the edges of connecting node pairs.

Github code can be found here.