Combined 2D–3D categorization and classification for multimodal perception systems

Abstract

This article describes an object perception system for autonomous robots performing everyday manipulation tasks in kitchen environments. The perception system gains its strengths by exploiting that the robots are to perform the same kinds of tasks with the same objects over and over again. It does so by learning the object representations necessary for the recognition and reconstruction in the context of pick and place tasks.The system employs a library of specialized perception routines that solve different, well-defined perceptual sub-tasks and can be combined into composite perceptual activities including the construction of an object model database, multi-modal object classification, and object model reconstruction for grasping.We evaluate the effectiveness of our methods, and give examples of application scenarios using our personal robotic assistants acting in a human living environment.

Object Categorization in Clutter Using Additive Features and Hashing of Part-Graph Descriptors

“…This allows for cheap partial re-training when novel objects need to be added, as the training set is partitioned into multiple parts, and separate classifiers. This is in contrast with our previous work [3], but obviously assumes known labels of the novel exemplars.…”

“…This was the case in our earlier experiments based on a smaller laser-based dataset [22], where this strategy improved classification rates by 5%. Since we are dealing with tabletop scenes, the supporting plane can be removed prior to processing, and only points above it considered as in [3]. Small segments are discarded, and for each segment we subsequently compute the GRSD-feature (more detailed description in section 6) and store it for later use.…”

“…In order to match the perception capabilities of humans, Dickinson in [2] advocates that searching for predefined templates is not enough, and that recognition of new exemplars of known categories has to be facilitated. On this premise, Marton et al [3] use geometric cues for categorization and visual cues for instance classification.…”

“…Since teaching all the possible objects a robot might encounter is not always possible, categorization becomes an important step towards learning these novel objects [2]. Thus a mechanism is needed to enable a robot to autonomously acquire new object models as it operates in an environment, and to efficiently add it to its classification framework [3].…”

Comparing and combiningof point cloud descriptors features in static hand posture recognition

“…GRSD opisuje radialne relacje punktów z ich sąsiedztwem. Aby zrozumieć jego działanie, należy najpierw przedstawić lokalny deskryptor RSD, który został opisany w pracach [8], [9].…”