Can we unify monocular detectors for autonomous driving by using the pixel-wise semantic segmentation of CNNs? | Publicación