DMS Graduate Student Seminar

Speaker: Dr. Luke Oeding

Title: Images, Mathematics, Machine Vision, and Neural Networks 

Abstract: In the 1960s computers were in their infancy. Computer scientists were looking to turn everything into 0’s and 1’s. Today the digital images we see on our devices today are no more than large matrices of integers rendered on a screen. Like the scene from The Matrix, our entire world seems to be an endless stream of digits flowing by. Back in the 60’s American painter and abstract impressionist Jasper Johns took the opposite approach, and instead of digitizing his painting, he painted the digits themselves in his series of works works entitled «0 through 9.» Perhaps inspired by arithmetic operations on matrices (images) he used his canvas and bush to explore the geometric visualization of operations on digits. We will explore variations on the mathematical themes in Johns’ work, starting from the most basic operations on images, producing our own versions of 0-9. Then we will discuss more modern mathematics and computer science, including some current areas of research, such as convolutional filtering, feature detection, augmented reality, and neural networks.