"A Primer on Image Segmentation"
Image segmentation is an example of a task that is remarkably easy for humans but incredibly hard for computers. I will survey several of the most common mathematical approaches to image segmentation, including local thresholding, edge-detection methods, PDE methods, and neural networks. For each, I will describe cases where these methods work and also their limitations, focusing on medical applications relating to both previous and proposed research I have conducted. Time permitting, I will include a discussion of the computational effort required in relation to each of these methods, and discuss their scalability to 3D medical image problems.