Bayesian Eigenobjects: A Unified Framework for 3D Robot Perception
نویسندگان
چکیده
We introduce Bayesian Eigenobjects (BEOs), a novel object representation that is the first technique able to perform joint classification, pose estimation, and 3D geometric completion on previously unencountered and partially observed query objects. BEOs employ Variational Bayesian Principal Component Analysis (VBPCA) directly on 3D object representations to create generative and compact probabilistic models for classes of 3D objects. Using only depth information, we significantly outperform the current state-of-the-art method for joint classification and 3D completion in both accuracy and query time. Additionally, we show that BEOs are well suited for the extremely challenging task of joint classification, completion, and pose estimation on a large dataset of household objects.
منابع مشابه
Generalized 3D Object Representation using Bayesian Eigenobjects
Current methods for representing 3D objects for robotic interaction have significant limitations. They do not allow knowledge transfer from previously encountered objects to similar but novel objects, they construct large object databases that do not scale and are expensive to query, or they require hand tuned object models. We propose the use of Variational Bayesian Principal Component Analysi...
متن کاملHierarchical Log-Spherical Inference Grid – An Unconventional Approach to Robotic Perception and Action
In this text, an unconventional paradigm for robotic multisensory perception and action will be presented in the form of a generalisation of a framework devised in previous work by the authors, the Bayesian Volumetric Map (BVM). The BVM, a log-spherical inference grid providing an egocentric and probabilistic representation of spatial information, was designed to bridge multisensory perception ...
متن کاملA Bayesian Binaural System for 3D Sound-Source Localisation
In this text we present a Bayesian system of auditory localisation in distance, azimuth and elevation using binaural cues only. We describe its supporting sensor model and calibration procedure. The binaural system is also integrated in a spatial representation framework for multimodal perception of 3D structure and motion — the Bayesian Volumetric Map (BVM). This solution will enable the imple...
متن کاملMultimodal Active Exploration Using a Bayesian Approach
In this text, we use a Bayesian framework for active multimodal perception of 3D structure and motion — which, while not strictly neuromimetic, finds its roots in the role of the dorsal perceptual pathway of the human brain — to implement a strategy of active exploration based on entropy. The computational models described in this text support a robotic implementation of multimodal active perce...
متن کاملLearning emergent behaviours for a hierarchical Bayesian 3 framework for active robotic perception 4 João
8 Abstract In this research work, we contribute with a 9 behaviour learning process for a hierarchical Bayesian 10 framework for multimodal active perception, devised to be 11 emergent, scalable and adaptive. This framework is com12 posed by models built upon a common spatial configura13 tion for encoding perception and action that is naturally 14 fitting for the integration of readings from mu...
متن کامل