Development of convolutional neural networks for an electron-tracking Compton camera

Electron-tracking Compton camera, which is a complete camera with tracking scattering electron by gas micro time projection chamber, expected to open up MeV gamma-ray astronomy. The technical challenge for achieving several degrees of the point spread function precise determination electron-recoil direction and position from track images. We attempted reconstruct these parameters using convolutional neural networks. Two network models were designed predict recoil position. These marked 41$~$degrees angular resolution 2.1$~$mm 75$~$keV simulation data in Argon-based at 2$~$atm pressure. In addition, ETCC was improved 15$~$degrees 22$~$degrees experimental 662$~$keV source. performances greatly surpassed that traditional analysis.