Browse Prior Art Database

ESTABLISHING CORRESPONDENCE OF NON-RIGID OBJECTS USING SMOOTHNESS OF MOTION

IP.com Disclosure Number: IPCOM000128471D
Original Publication Date: 1984-Jan-01
Included in the Prior Art Database: 2005-Sep-16
Document File: 9 page(s) / 105K

Publishing Venue

Software Patent Institute

Related People

Jain, Ramesh: AUTHOR [+4]

Abstract

Identifying the same physical object in more than one image, the object correspondence problem, is important in many applications. Most approaches for establishing correspondence in a sequence of frames depend heavily on the 2-D features and 2-D distance in the location of correspondence tokens in two frames. By using a sequence of frames, rather than just two or three frames, it is possible to exploit the fact that due to inertia, the motion of an object can not change instantaneously. We use smoothness of motion of an object for solving the correspondence problem for non-rigid motions. By using path coherence, the correspondence problem is formulated as an optimization problem. This paper presents our initial study of the efficacy of the proposed approach. Index Terms: Correspondence, dynamic scene analysis, structure from motion, path coherence, smoothness of motion.

This text was extracted from a PDF file.
This is the abbreviated version, containing approximately 11% of the total text.

Page 1 of 9

THIS DOCUMENT IS AN APPROXIMATE REPRESENTATION OF THE ORIGINAL.

ESTABLISHING CORRESPONDENCE OF NON-RIGID OBJECTS USING SMOOTHNESS OF MOTION

Ramesh Jain and I. K. Sethi

CRL-TR-10-84

THE UNIVERSITY OF MICHIGAN COMPUTING RESEARCH LABORATORY1

JANUARY 1984

Room 1079, East Engineering Building

Ann Arbor, Michigan 48109
USA
Tel: (313) 763-8000

Establishing Correspondence of Non-Rigid Objects Using Smoothness of Motion Ramesh Jain2 Electrical and Computer Engineering The University of Michigan
Ann Arbor, MI 48105 and
I. K. Sethi

Department of Computer Science
Wayne State University
Detroit, MI 48202

Abstract

Identifying the same physical object in more than one image, the object correspondence problem, is important in many applications. Most approaches for establishing correspondence in a sequence of frames depend heavily on the 2-D features and 2-D distance in the location of correspondence tokens in two frames. By using a sequence of frames, rather than just two or three frames, it is possible to exploit the fact that due to inertia, the motion of an object can not change instantaneously. We use smoothness of motion of an object for solving the correspondence problem for non-rigid motions. By using path coherence, the correspondence problem is formulated as an optimization problem. This paper presents our initial study of the efficacy of the proposed approach.

Index Terms: Correspondence, dynamic scene analysis, structure from motion, path coherence, smoothness of motion.

1. Introduction

1 Any opinions, findings, and conclusions or recommendations expressed in this publication are those of the authors.

2 Mail all correspondence to this author.

University of Michigan Computing Research Laboratory Page 1 Jan 01, 1984

Page 2 of 9

ESTABLISHING CORRESPONDENCE OF NON-RIGID OBJECTS USING SMOOTHNESS OF MOTION

Structure from motion has attracted significant research efforts in recently 3 from researchers working in the field of dynamic scene analysis4. Ullman introduced the rigidity assumption which states that any set of elements undergoing a 2-D transformation which has a unique interpretation as a rigid body moving in space should be so interpreted. The research in human perception 5 suggests that the human visual system exploits this fact. The rigidity assumption has been very influential in structure from motion research. The research in the recovery of 3-D structure from image sequences may be divided in two general classes. Some researchers have been concerned with the problem of recovery of the structure and motion using a minimal number of points in a minimal number of frames. Recently, the trajectory based recovery has attracted some attention.

1.1. Recovering Structure Using Tokens

Suppose that we apply an interest operator to consecutive frames of a sequence and extract some interesting points or tokens, such as corners, then using some method, manage to succeed in solving the correspondence problem. Ullman shows that if token corres...