Tetris-Inspired Detector with Neural Network for Radiation Mapping

Published in Nature Communications, 2024

Paper / MIT News / Physicsworld / Nuclear Engineering International / GitHub /

A radiation mapping framework using Tetris-inspired detector pixels: inter-pixel padding plus neural networks trained on Monte Carlo data let a detector with as few as four pixels predict source direction, and a moving detector with maximum a posteriori estimation localizes sources, verified in field tests.

Recommended citation: Okabe, R.†, Xue, S.†, Vavrek, J.†, Yu, J., Pavlovsky, R., Negut, V., Quiter, B., Cates, J., Liu, T., Forget, B., Jegelka, S., Kohse, G., Hu, L-W., & Li, M. (2024). Tetris-Inspired Detector with Neural Network for Radiation Mapping. Nature Communications 15, 3061.
Download Paper