Speaker
Description
Clustering of calorimetric signals in the ATLAS detector has typically been performed using the topocluster algorithm, following cell signal-significance patterns. In this work we present a machine learning alternative to topoclustering. Using current topological cell clusters as indicators of physical significance we use a convolutional neural network (CNN) to identify regions of interest in the calorimeter. We introduce a novel data pre-processing pipeline transforming the ATLAS calorimeter into a two-dimensional representation in η,φ; building upon previous treatments of jets as images in particle physics. The performance of the object detection architecture, which targets real-time applications, is evaluated on a set of simulated particle interactions in the ATLAS detector.
Minioral | Yes |
---|---|
IEEE Member | No |
Are you a student? | Yes |