HIGHLIGHTS
SUMMARY
In these approaches, domain knowledge on wave scattering physics is incorporated into the input or in the internal architecture of the neural_network, having a beneficial effect on its performance. In this paper, the problem of retrieving the morphological and electromagnetic properties of unknown targets encoded in the OSM images is cast in terms of a segmentation task performed by a U-Net architecture. The U-Net is trained to associate the pixels of the OSM imaging results with a finite set of contrast values. Conversely, the capability of supervised machine_learning to build models to . . .
If you want to have access to all the content you need to log in!
Thanks :)
If you don't have an account, you can create one here.