Splitted Neural Networks for Equipment Inspection: Experiments and Resources

Luan Assis Gonçalves, Tiago Guerreiro, Samya Pinheiro, Flavio Mendes de Brito, Ingrid Nascimento, Neiva Linder, Aldebaro Klautau

DOI: 10.14209/SBRT.2020.1570661611

Evento: XXXVIII Simpósio Brasileiro de Telecomunicações e Processamento de Sinais (SBrT2020)

Keywords: splitted neural network networked environment equipment inspection semantic segmentation

Abstract

This paper presents a feasibility study of performing equipment inspection on networked environments by exploring networks specially designed for semantic segmentation task. It is assumed that the neural network needs to be split and the resulting two pieces need to be allocated in two devices. The provided results suggest that the current state of the art on semantic segmentation is not well-suited for the splitted networks applications even in the context of the 5G networks.

Download