ND Simulation/Data full semantic segmentation, which is capable of identifying tracks/showers/blips and to some degree, the particle and physics labels for each of those. #51
Need to add in vertex/track/shower end points into the segmentation. There are some ideas on how to do this that we can try, such as adding a label which is a Gaussian heat map centered on the pixel points, and/or the value of the distance to the nearest point.
For the heat map, we would need to define a sigma value. This is essentially "keypoint estimation"
Need to add in vertex/track/shower end points into the segmentation. There are some ideas on how to do this that we can try, such as adding a label which is a Gaussian heat map centered on the pixel points, and/or the value of the distance to the nearest point.
For the heat map, we would need to define a sigma value. This is essentially "keypoint estimation"