Evaluation of the Vocabulary Tree Object Recognition Method For a Mobile Robotic Platform

This paper addresses object perception applied to mobile robotics. Being able to perceive semantically meaningful objects in unstructured environments is a key capability in order to make robots suitable to perform high-level tasks in home environments. However, finding a solution for this task is daunting: it requires the ability to handle the variability in image formation in a moving camera with tight time constraints. Here we extensively evaluate the vocabulary tree object recognition method in a very challenging dataset of 30 objects captured while navigating with a mobile robot in a typical indoor setting. In order to make the method suitable for this detection task, a sliding windows method as well as two segmentation methods, based in intensity and depth information, are evaluated. The obtained results are compared with those obtained in the same dataset by two other popular state of the art object recognition methods: SIFT object detection method and the Boosted Cascade of Weak Classifiers.