|By combining techniques from geometric hashing and structural indexing, we have developed a new representation for recognition of free-form objects from three dimensional data. The representation comprises descriptive spin-images associated with each oriented point on the surface of an object. Constructed using single point bases, spin-images are data level shape descriptions that are used for efficient matching of oriented points. During recognition, scene spin-images are indexed into a stack of model images to establish point correspondences between a model object and scene data. Given oriented point correspondences, a rigid transformation that maps the model into the scene is calculated and then refined and verified using a modified iterative closest point algorithm.
Indexing of oriented points bridges the gap between recognition by global properties and feature bases recognition without resorting to error-prone segmentation or feature extraction. It requires no knowledge of the initial transformation between model and scene, and it can register fully 3-D data sets as well as recognize objects from partial views with occlusions. We present results showing simultaneous recognition of multiple 3-D anatomical models in range images and range image registration in the context of interior modeling of an industrial facility.
Sponsor: Dept. of Energy
Grant ID: DE-AC21-92MC29104
Associated Center(s) / Consortia: Vision and Autonomous Systems Center
Number of pages: 36
|Andrew Johnson and Martial Hebert, "Recognizing Objects by Matching Oriented Points," tech. report CMU-RI-TR-96-04, Robotics Institute, Carnegie Mellon University, May, 1996|
author = "Andrew Johnson and Martial Hebert",
title = "Recognizing Objects by Matching Oriented Points",
booktitle = "",
institution = "Robotics Institute",
month = "May",
year = "1996",
address= "Pittsburgh, PA",
|The Robotics Institute is part of the School of Computer Science, Carnegie Mellon University.|
Contact Us | Update Instructions