On 05.01.2023 at 11:00 a.m. the 18th lecture of the Living Lab lecture series took place. In this talk, ScaDS.AI scientific researcher Jannik Irmai talked about Graph Multi-Cuts and Image Analysis. Add this event to your calendar (iCal).
Graph Multi-Cuts and Image Analysis
In this lecture, we learn how graph theory and combinatorial optimization can be used to solve problems from the field of image analysis. We consider the challenging task of segmenting a three dimensional volume image of an electron microscopy scan of a mouse neocortex. The state of the art method for solving this segmentation problem consists of a convolutional neural network and a post-processing step, which involves solving the graph multi-cut problem. We will see how segmenting an image can be understood as a problem of cutting a graph into multiple components. The optimal multi-cut can be found by means of integer linear programming. This method, however, does not scale to large instances. To address this problem, we present a second algorithm that finds a good but not necessarily optimal solution efficiently. Finally, we will see that the multi-cut problem and algorithms have further applications beyond image segmentation.
Missed this Living Lab lecture on Graph Multi-Cuts and Image Analysis?
You can rewatch this lecture on YouTube.
Living Lab Lecture Series
The Living Lab Lecture Series gives you an in-depth insight into the many research topics of ScaDS.AI Dresden/Leipzig. From Natural Language Processing to Ethics and Moral Code in AI, a great variety of topics are discussed. You can join our lectures every first thursday of the month or watch them on YouTube afterwards. If you have ideas for topics to discuss in the future, please let our Living Lab team know. We suggest for you to regularly check our event calendar, to never miss out on upcoming lectures or other interesting events organized by or in cooperation with our center.
Find out more about our Living Lab Lecture Series!