Browse Prior Art Database

Combining Motion and Contrast for Segmentation

IP.com Disclosure Number: IPCOM000128564D
Original Publication Date: 1979-Dec-31
Included in the Prior Art Database: 2005-Sep-16
Document File: 9 page(s) / 35K

Publishing Venue

Software Patent Institute

Related People

William Bo Thompson: AUTHOR [+3]

Abstract

A method is presented for partitioning a scene into regions corresponding to surfaces with distinct velo-cities. Both motion and contrast information are ,incorporated into the segmentation process. Velocity estimates for each point in a scene are obtained using a local, non-matching technique not dependent on any prior boundary analysis. The actual segmenta-tion is accomplished using a region merging procedure which combines regions based on similarities in both brightness and motion. The method is effective in determining object boundaries not easily found using analysis applied only to a single image frame.

This text was extracted from a PDF file.
This is the abbreviated version, containing approximately 10% of the total text.

Page 1 of 9

THIS DOCUMENT IS AN APPROXIMATE REPRESENTATION OF THE ORIGINAL.

Combining Motion and Contrast for Segmentation

by William Bo Thompson

Computer Science Department

136 Lind Hall Institute of Technology University of Minnesota Minneapolis, Minnesota 55455 Cover design courtesy of Ruth and Jay Leavitt

Technical Report 79-7 March 1979 rA Adt. ~fo Combining Motion and Contrast for Segmentation

William B. Thompson Computer Science Department University of Minnesota Minneapolis, MN 55455

ABSTRACT

A method is presented for partitioning a scene into regions corresponding to surfaces with distinct velo-cities. Both motion and contrast information are ,incorporated into the segmentation process. Velocity estimates for each point in a scene are obtained using a local, non-matching technique not dependent on any prior boundary analysis. The actual segmenta-tion is accomplished using a region merging procedure which combines regions based on similarities in both brightness and motion. The method is effective in determining object boundaries not easily found using analysis applied only to a single image frame.

I. INTRODUCTION

Motion is a crucial property of many visual environments. Knowledge of object velocities and trajectories is clearly impor-tant for scene interpretation. Motion is also useful as a cue for scene segmentation. Velocity information may be used to link adjacent but visually dissimilar surfaces or to divide surfaces not easily separable by static criteria alone. Often, ambiguous object boundaries in a single image frame are easily resolved when dynamic effects are evaluated based on a sequence of frames. This paper describes a technique for combining both motion and brightness information into a single segmentation procedure which determines the boundaries of moving objects. The most straightforward approach to locating moving sur-faces is to compare a sequence of accurately registered image frames searching for areas of change. If velocity estimation is not required, simple pixel-at-a-time subtractive techniques applied to two frames may be appropriate [1]. In order to deter-mine speed and direction of moving objects, differences must be evaluated over longer sequences [2]. In fact, if the actual boun-daries of moving objects are required, these objects must be observed over an interval that is long enough for the objects to travel at least as far as their length along the direction of motion. A second approach involves tracking identifiable image structures from frame to frame [3,4,5]. These systems are capable of accurately estimating surface translation, but usually only for a relatively sparse sampling of points. Potter has described a similar system designed to estimate velocity at every point in a scene and thus allow for segmentation~based on motion [6]. At each point in one frame, a skeletal template is defined. This template is then searched for in a second

University of Minnesota Page 1 Dec 31, 1979

Page 2 of 9

Combin...