@misc{oai:repo.qst.go.jp:00083746,
 author = {Tanabe, Hiroyuki and Takafumi, Asai and Masato, Kanasaki and Satoshi, Jinno and Nobuko, Kitagawa and Tomoya, Yamauchi and Kunihiro, Morishima and Yuji, Fukuda and Tanabe, Hiroyuki and Takafumi, Asai and Yuji, Fukuda},
 month = {Nov},
 note = {In the interaction between intense laser and the target matter, near-100-MeV proton acceleration is demonstrated. It offers that the potential to realize a energy saving compact particle accelerators in the future. To understand the acceleration process mechanism, precise measurement of both the energy spectrum and the two-dimensional distribution is required. Against this background, we have developed a new measurement method for laser-accelerated sub-GeV-class protons using the nuclear emulsion. Based on the Multiple Coulomb Scattering (MCS) method in an Emulsion Cloud Chamber (ECC), which is a stack of nuclear emulsion films and scatterer plates, the incident energies were inversely evaluated by the scattering angle. The proof-of-principle simulation has been conducted with GEANT-4 Monte Carlo code. To analyze the proton tracks, we have applied the deep learning technique to obtain the incident energy from the amount of MCS in each layer of nuclear emulsion. The median of energy determination coefficient is 0.73 with ramp activation function. The coefficients of the present studies are equal or smaller than the conventional regression model, and we are trying to improve the determination coefficient by optimization of the calculation conditions., ICMaSS2021},
 title = {Analysis Method of Laser-accelerated Sub-GeV-class Proton Tracks in Emulsion Cloud Chamber using Deep Learning Technique},
 year = {2021}
}