“…At present, there are methods that produce impressive results in a wide variety of challenging scenarios corrupted by lighting changes, cluttered backgrounds, partial occlusions, viewpoint and scale changes, and large intra-class variations [4,15,25,33,45,48,51] This progress in object recognition has had a positive impact in many application fields such as robotics, where computer vision algorithms have been used for diverse robotics tasks such as object recognition and grasping [3,5], detection and tracking of people in urban settings [10,34,39], human-robot interaction [40,47], and robot localization and navigation [11,16,26].…”