Fig. 14From: Fully convolutional networks for structural health monitoring through multivariate time series classificationDamage detection, case 1. Training and validation of the two branches convolutional architecture: evolution of the loss \(J_d \left( {\varvec{Y}}, {\varvec{p}} \right) \) on \({\mathbb {D}}^d_{train}\) and \({\mathbb {D}}^d_{val}\) (left column), and of \({\mathcal {G}}_d\) accuracy \(A_d\) (right column) on \({\mathbb {D}}^d_{train}\) and on \({\mathbb {D}}^d_{val}\), both for the SNR\(=15\) dB case (top row) and for the SNR\(=10\) dB case (bottom row)Back to article page