Object Recognition By Functional Parts
.
Object Recognition by Functional Parts.
In IUW, 1994
Online Version
A pdf version is available for download.
Abstract
We present an approach to function-based object recognition that reasons about the functionality of an object's intuitive parts. We extend the popular "recognition by parts" shape recognition framework to support "recognition by functional parts", by combining a set of functional primitives and their relations with a set of abstract volumetric shape primitives and their relations. Previous approaches have relied on more global object features, often ignoring the problem of object segmentation and thereby restricting themselves to range images of un-occluded scenes. We show how these shape primitives and relations can be easily recovered from super quadric ellipsoids which, in turn, can be recovered from either range or intensity images of occluded scenes. Furthermore, the proposed framework supports both unexpected (bottom-up) object recognition and expected (top-down) object recognition. We demonstrate the approach on a simple domain by recognizing a restricted class of hand-tools from 2-D images.
Keywords
Co-authors
Bibtex Entry
@inproceedings{RivlinDR94i-o,
title = {Object Recognition by Functional Parts},
author = {Ehud Rivlin and Sven J. Dickinson and Azriel Rosenfeld},
year = {1994},
booktitle = {IUW},
keywords = {Function},
abstract = {We present an approach to function-based object recognition that reasons about the functionality of an object's intuitive parts. We extend the popular "recognition by parts" shape recognition framework to support "recognition by functional parts", by combining a set of functional primitives and their relations with a set of abstract volumetric shape primitives and their relations. Previous approaches have relied on more global object features, often ignoring the problem of object segmentation and thereby restricting themselves to range images of un-occluded scenes. We show how these shape primitives and relations can be easily recovered from super quadric ellipsoids which, in turn, can be recovered from either range or intensity images of occluded scenes. Furthermore, the proposed framework supports both unexpected (bottom-up) object recognition and expected (top-down) object recognition. We demonstrate the approach on a simple domain by recognizing a restricted class of hand-tools from 2-D images.}
}