“…Furthermore, NN1 is replaced with an untrained three-layer FCN(named the NN3), and the NN2 structure remains unchanged Table 2. The relevant parameters of the previous model Channels [100, 50, 20, 3] Kernel size [(5,5), (3,3), (3,3), (2,2)] Max pooling [2, 2, 2, 2] NN3 layer size [64, 64, 3]NN2 layer size[16,1]…”