TPC track denoising and recognition using convolutional neural networks
Matěj Gajdoš
The Faculty of Mathematics and Physics, Charles University; IEAP, CTU in Prague
Abstrakt:
Spurious signals caused by microdischarges are a known effect inherent to all gaseous detectors, such as time projection chambers (TPC). During the reconstruction in imaging and tracking detectors, these signals are added to the actual track-generated signal as extra pixels or clusters, compromising the performance of the detector. In order to denoise these artifacts, we leveraged an existing technique, which uses 2D convolutional neural networks, to denoise TPC events represented by 3D arrays. In this seminar, I will guide you through this process, discussing the emerging challenges when real data are used instead of toy examples with clean datasets.
Seminář se koná v úterý 6. května ve 14:00
v zasedací místnosti ÚTEF ČVUT, Praha 1, Husova 240/5.
| Ing. Bartoloměj Biskup, Ph.D. tajemník semináře |
doc. Ing. Ivan Štekl, CSc. ředitel ÚTEF |
doc. Dr. André Sopczak předseda NPS, ČS IEEE |
NUCLEAR & PLASMA SCIENCES SOCIETY CHAPTER
IEEE Czechoslovakia section
https://www.ieee.cz/main/section/nps/