AUTHOR=Yisheng Hao , Zhen Wu , Yanheng Pu , Yuhang Zhang , Rui Qiu , Hui Zhang , Junli Li 

TITLE=Validation of the neural network for 3D photon radiation field reconstruction under various source distributions

JOURNAL=Frontiers in Energy Research

VOLUME=Volume 11 - 2023

YEAR=2023

URL=https://www.frontiersin.org/journals/energy-research/articles/10.3389/fenrg.2023.1151364

DOI=10.3389/fenrg.2023.1151364

ISSN=2296-598X

ABSTRACT=In practice, the radiation parameters in a radiation space can only be predicted by some detectors. In this paper, a five-layer fully connected neural network is proposed based on typical 3D geometric examples, which utilize the gamma flux of individual positions that simulate the detector position in the actual situation as the input to subsequently train and test the gamma radiation field in 3D space under different source term distributions. Finally, some new examples are predicted with the obtained neural network. The results show that the mean percentage change error (PCT) for the test set under different source term distributions ranges from 0.53% to 3.11%, with a maximum error value of 3.66%. This is within the error range of the given value of the measurement input (±10%). The results further show that the neural network method can be used for the direct reconstruction of the 3D radiation field with some simple source terms. This method has distinct advantages and practical value in real operations within radiation spaces.