Method for automatically locating image pattern in digital images

• US 6,249,590 B1
• Filed: 02/01/1999
• Issued: 06/19/2001
• Est. Priority Date: 02/01/1999
• Status: Expired due to Fees

• Current Assignee: CareStream Health Incorporated
• Original Assignee: Eastman Kodak Company
• Inventors: Lee, Hsien-Che, Young, Susan S.
• Primary Examiner(s): Mancuso, Joseph
• Assistant Examiner(s): Tabatabai, Abolfazl
• Application Number:
  US09/243,400
• Time in Patent Office: 869 Days
• Field of Search: 382/103, 382/28, 382/133, 356/401, 701/301, 348/169
• US Class Current: 382/103
• CPC Class Codes: A61B 6/505 for diagnosis of bone
• CPC Class Codes: G06T 2207/30008 Bone
• CPC Class Codes: G06T 7/70 Determining position or orientation of objects or cameras camera calibration G06T7/80

Abstract

A method for automatically locating instances of a target pattern in digital images, comprising the steps of: providing a digital image; detecting a first simple feature associated with the target patterns in the digital image; for each detected feature, searching, in its spatial neighborhood, a second or a plural of other features associated with the target patterns; for each pair of plural of features detected, checking the consistency of image intensity profile with the target pattern within the spatial neighborhood as delimited by the feature points detected in the searching and detecting steps; and labeling those image regions that are found to be consistent with the structure of the target pattern in the checking step.

Claims

1. A method of automatically locating intervertebral disks of lumbar spine in digital images comprising the steps of:

providing a digital image having intervertebral disks of lumbar spine;

applying a set of directional ridge filters;

searching for a pair of disk space ridge points;

computing the image pattern histogram;

constructing the tone scale curve for the provided digital image from the image pattern histogram; and

applying the tone scale curve to said provided digital image to produce an output image.

2. The method of claim 1 wherein said directional ridge filter is designed by:

determining a horizontal ridge filter which is the waveform function in x-direction and a constant in y-direction;

determining a directional ridge filter by rotating the said horizontal ridge filter with certain increments in plural directions; and

determining the said ridge filter parameters based on the physical pixel spacing size.

3. The method of claim 1 wherein said searching step includes:

generating a line along the maximum response direction of the said ridge filter;

searching for another ridge point on the said line (in both directions);

locating an ending ridge point by determining the absolute difference of the angular directions of the starting and the ending points below a certain threshold;

determining a profile vector of the line connecting the starting and the ending ridge points;

checking the local maxima conditions, that is, for a valid pair of disk space ridge points, the starting and the ending ridge points should be local maxima of the said profile vector;

checking the local minimal conditions, that is, for a valid pair of disk space ridge points, the minimum of the said profile does not occur at the boundary (ridge points);

checking the range condition 1 of the boundary ridge points, that is, for a valid pair of disk space ridge points, the maximum of the differential boundary changes, (differential boundary change is defined as the difference of the code value of one ridge point and the point next to it), should be a fraction of the code value range of the said profile vector;

checking the range condition 2 of the boundary ridge points, that is, for a valid pair of disk space ridge points, the differential change of ridge points, which is defined as the difference of the code values of two ridge points (the starting and the ending ridge points), should be a fraction of the code value range of the said profile vector; and

checking the convex condition, that is, for a valid pair of disk space ridge points, the points between the starting and the ending ridge points are below the line that connects the two ridge points on the said profile.

4. The method of claim 1, wherein said image pattern histogram is constructed from the searched valid disk space ridge points and the points between the said profile vector.