Methods and apparatuses for analyzing images

• US 7,885,438 B2
• Filed: 09/20/2002
• Issued: 02/08/2011
• Est. Priority Date: 02/12/1997
• Status: Expired due to Fees

• Current Assignee: University of Iowa Research Foundation
• Original Assignee: University of Iowa Research Foundation
• Inventors: Uppaluri, Renuka, Hoffman, Eric A., McLennan, Geoffrey, Mitsa, Theophano, Sonka, Milan
• Primary Examiner(s): Le; Brian Q
• Application Number:
  US10/251,446
• Publication Number:
  US 20030103665A1
• Time in Patent Office: 3,063 Days
• Field of Search: 382128-133, 382164-165, 382168-170, 382/173, 382190-191, 382/192, 382/194, 382/210, 382/224, 382/249, 382/257, 382/218, 128/922, 250/363.04, 356/39, 377/10, 345/423
• US Class Current: 382/128
• CPC Class Codes: G06T 7/0012 Biomedical image inspection

Abstract

A method for automated analysis of textural differences present on an image volume. The image volume includes a plurality of volume elements, and each volume element has a gray level. The method includes defining a volume of interest (VOI); performing texture measures within the VOI; and classifying the VOI as belonging to a tissue pathology class based upon the texture measures. Computer readable media encoded with computer readable instructions for carrying out these functions. An apparatus that includes an image input adapted to receive a diagnostic medical image. The image includes a plurality of pixels, and each pixel has a particular gray level. The apparatus also includes a display for displaying a graphical user interface and the received image; and a processor adapted to perform texture measures on one or more groups of pixels within the image and classify each group of pixels to a tissue pathology class based upon the textures measures. The processor is further adapted to (1) associate a color to each group of pixels indicative of the group'"'"'s tissue pathology class, (2) cause the display to display one or more of the colors on the image at the location of the associated group or groups of pixels, (3) permit a user to manually associate a tissue pathology class to a group of pixels, and (4) cause the display to display the manually-associated tissue pathology class.

Claims

1. A method for automated analysis of textural differences present on an image volume, the image volume comprising a plurality of volume elements, each volume element having a gray level, the method comprising:

using a computer to perform the following steps;

removing structures within the image volume by assigning a particular gray level to the volume elements that form the structures;

after the removing, defining a volume of interest (VOI) from which at least one of the structures has been removed;

displaying the VOI;

performing texture measures within the VOI; and

classifying the VOI as belonging to a tissue pathology class based upon the texture measures.

2. The method of claim 1, where the performing comprises:

performing at least one first order texture measure within the VOI to describe a frequency of occurrence of all gray levels assigned to volume elements of the VOI; and

performing at least one second order texture measure within the VOI to describe spatial interdependencies between the volume elements of the VOI.

3. The method of claim 2, where prior to the classifying, the method further comprises:

eliminating the first order and second order texture measures that are redundant or fail to properly distinguish a particular tissue pathology class.

4. The method of claim 1, where prior to defining the VOI, the method further comprises:

forming volume element regions within the image volume by assigning volume elements located adjacent one another a common gray level provided the adjacent volume elements'"'"' gray levels differ by an insignificant amount.

5. The method of claim 4, where each commonly assigned gray level is the average of the adjacent volume elements'"'"' gray levels that differ in value by an insignificant amount.

6. The method of claim 5, where the adjacent volume elements'"'"' gray levels differ by an insignificant amount is a difference in gray level of 20 or less.

7. The method of claim 1, further comprising:

determining the number of times a particular gray level “

i”

of a volume element “

x”

within the VOI and a particular gray level “

j”

of another volume element within the VOI are separated by a distance of approximately two pixels in a direction “

N”

relative to volume element “

x”

.

8. The method of claim 1, where prior to the classifying, the method further comprises:

centering a volume element block about each volume element of the VOI;

determining the average of absolute gray level intensity differences of each possible volume element-pair separated by a distance “

d”

within each volume element block; and

assigning the volume element, about which each volume element block is centered, a stochastic fractal value based upon the average absolute gray level intensity differences obtained.

9. The method of claim 8 further comprising:

assigning the volume elements of the VOI only one of two binary values dependent upon their respective gray levels;

mapping each image onto a grid of super-volume elements of increasing size “

e”

;

determining the number of the super-volume elements that are one binary value within the VOI; and

determining a geometric fractal value of the VOI based upon the determined number of super-volume elements that are the one binary value.

10. The method of claim 1, where prior to the classifying, the method further comprises:

providing known samples of particular tissue pathology classes;

performing first order and second order texture measures from the samples; and

storing the first order and second order texture measures obtained from the samples.

11. The method of claim 10 further comprising:

determining the probability that the VOI belongs to a particular tissue pathology class based upon the stored texture measures and the texture measures performed on the VOI; and

the classifying comprises classifying the VOI to the particular tissue pathology class for which the probability is the highest.

12. The method of claim 1, where the performing comprises performing texture measures on one or more groups of volume elements within the VOI, the classifying comprises classifying the VOI as belonging to a tissue pathology class based upon the texture measures, and where the method further comprises:

assigning a color code to the VOI indicative of the tissue pathology class of the VOI.