Extracting Signal Electron Trajectories in the COMET Phase-I Cylindrical Drift Chamber Using Deep Learning

Authors

Fumihiro Kaneko 1,*, Yoshitaka Kuno 2,3,†, Joe Sato 4,‡, Ikuya Sato5,§,Dorian Pieters3,¶, and Chen Wu 2,6,7,∥

Affiliations

¹Solutions,PreferredNetworks,Inc.,1-6-1Otemachi,Tokyo100-0004,Japan
²ResearchCenterof NuclearPhysics,OsakaUniversity,10-1Mihogaoka,Ibaraki,Osaka565-0871, Japan
³Departmentof Physics,Facultyof Science,OsakaUniversity,1-1Machikaneyama,Toyonaka,Osaka 560-0043,Japan
⁴Departmentof Mathematics,Physics,ElectricalEngineeringandComputerScience,Yokohama NationalUniversity,79-5Tokiwadai,Hodogaya-ku,Yokohama240-8501,Japan
⁵Departmentof Physics,Facultyof Science,SaitamaUniversity,255Shimo-Okubo,Sakura-ku,Saitama 338-8570,Japan
⁶Instituteof HighEnergyPhysics,ChineseAcademyof Sciences,19BYuquanRoad,Shijingshan District,Beijing100049,China
⁷SpallationNeutronSourceScienceCenter,Divisionof AcceleratorTechnology,1ZhongziyuanRoad, Dalang,Dongguan523803,China

Journal

Published Date

4/2/25

DOI

Abstract

We present a pioneering approach to tracking analysis within the Coherent Muon to Electron Transition (COMET) Phase-I experiment, which aims to search for the charged lepton flavor violating conversion process in a muonic atom, at J-PARC, Japan. This paper specifically introduces the extraction of signal electron trajectories in the COMET Phase-I cylindrical drift chamber (CDC) amid a high background hit rate, with more than 40% occupancy of the total CDC cells, utilizing deep learning techniques of semantic segmentation. Our model achieved remarkable results, with a purity rate of 98% and a retention rate of 90% for CDC cells with signal hits, surpassing the design-goal performance of 90% for both metrics. This study marks the initial application of deep learning to COMET tracking, paving the way for more advanced techniques in future research.