Dismiss
InnovationQ will be updated on Sunday, Oct. 22, from 10am ET - noon. You may experience brief service interruptions during that time.
Browse Prior Art Database

Border to Border Distance Measure for Characterizing Image Objects

IP.com Disclosure Number: IPCOM000045311D
Original Publication Date: 1983-Mar-01
Included in the Prior Art Database: 2005-Feb-06
Document File: 3 page(s) / 54K

Publishing Venue

IBM

Related People

Wahl, FM: AUTHOR

Abstract

This invention relates to a method for characterizing image objects using a border to border distance measure by encoding the measure in a two-pass run length procedure, which encoding preserves the border to border distance measure of each point.

This text was extracted from a PDF file.
At least one non-text object (such as an image or picture) has been suppressed.
This is the abbreviated version, containing approximately 58% of the total text.

Page 1 of 3

Border to Border Distance Measure for Characterizing Image Objects

This invention relates to a method for characterizing image objects using a border to border distance measure by encoding the measure in a two-pass run length procedure, which encoding preserves the border to border distance measure of each point.

Assume to be given a binary image with a subregion S which constitutes an object. The new distance d is a function of the two cartesian coordinates x, y ((x,y)EpsilonS) and an angle 0 measured from the x-axis (0 degrees less than or equal to 0 less than or equal to 180 degrees). It is defined as the length of a line segment B(1) B(2) with angle 0, which connects an inner point P(x,y) of S with two opposite border points B(1), B(2) of S, such that the line segment B(1) B(2) is entirely inside S (see Fig. 1). Note that for all tuples (x,y) lying on B(1)B(2), d is constant.

Two values of 0, namely those which minimize or maximize d respectively are of special interest. These two additional constraints result in the following distance mappings: D(min) (x,y)=min(d(x,y,0)) (1)

(0)

D(max) (x,y)=max(d(x,y,B)) (2)

(0).

In addition, an eccentricity mapping D(ecc) (x,y) can be defined by calculating the ratio D(max) (x,y) and D(min) (w,y) for any (x,y) inside S: D(ecc) (x,y)=max(d(x,y,0))/min(d(x,y,0)) (3).

With respect to an application in discrete space, it is important to find an efficient computation for the mappings proposed above. Fortunately, these quantities can be calculated very fast via the well-known run length algorithm, which here is applied in several directions to the discrete binary image data. However,...