Split-Et-Impera: A Framework for the Design of Distributed Deep Learning Applications | IEEE Conference Publication | IEEE Xplore