The University of Tartu and GScan combined machine learning models with muon trajectories and deflections to locate hidden objects more accurately in computer simulations than methods using only a single type of path. Their approach reduced the rate at which adjacent objects were incorrectly merged from about 15% to 2%. The study was published in IEEE Access, and the results currently apply only to simulations. Further testing is still needed under the more complex conditions of real buildings and bridges.
Source:Knowridge Science Report · Read original
Leave a Reply