Other results using this database presented by Tapia and Perez [37]; Ren and Li [27], not included in this table, were obtained removing many images, using only 7443 and 6840 samples out of 13, 233, respectively. It is important to underline that our results are obtained with the entire dataset, without removing any image. For this reason, both results cannot be compared on equal terms. On the other hand, very recently, Fazl-Ersi et al. [7] obtained a slightly better result in the Gallagher database. However, this result is obtained by using a complex ensemble composed by several handcrafted features such as LBPs, Color Histograms and SIFT. In contrast, our method is quite simple and generic to apply, so it may also work well in other computer vision tasks.