The exper- iment conducted on the test set by Dataset 2 [21] showed how the proposed autoencoded CNN model performs better than the other solutions for a few row–column repetitions (considering the range going from 1 to 3 repetitions). Then, the number of FP starts to become critical, ...