The shape of targets or a context of distracting objects did not affect accuracy or speed.
When we use direct imaging methods for solving inverse scattering problems, we observe artificial lines which make it hard to determine the shape of targets.
First, without the help of stereology, users need to explicitly provide the 3D shapes of target particles for synthesis.
Unlike bounding boxes used in ImageNet, our single model can draw the shapes of target objects, and in the meanwhile, classify the objects and calculate their sizes.
But occasionally, one of the muons will happen to hit an atomic nucleus, and when it does, it will change direction in a way that gives a clue about the shape of the target and the target's density.
Finally, the shape of the target object can be quantitatively obtained as follows: Δ L x, y, d = λ 2 π Δ ϕ x, y, d Δ n x, y, d (18).
It is also possible to analyze the size and the shape of each target object by observing its corresponding connected component.
The expected optimal window sizes depend on both the shape of the target's temporal profile, mainly on the target's peak width, which is inversely proportional to the target's velocity, and on the background scene, i.e., on the presence of clouds and their size and velocity, as stated in section "Window size".
The same arguments apply for localization and, especially, for shapes of targets.
