Method and apparatus for analyzing image structures

• US 6,173,083 B1
• Filed: 04/14/1998
• Issued: 01/09/2001
• Est. Priority Date: 04/14/1998
• Status: Expired due to Term

• Current Assignee: General Electric Company
• Original Assignee: General Electric Company
• Inventors: Avinash, Gopal B.
• Primary Examiner(s): ROGERS, SCOTT A
• Application Number:
  US09/059,886
• Time in Patent Office: 1,001 Days
• Field of Search: 382/131, 382/164, 382/168, 382/169, 382/171, 382/172, 382/132, 382/199, 382/173, 382/260, 382/261, 382/262, 382/266, 382/264, 382/194
• US Class Current: 382/260
• CPC Class Codes: G06T 2207/10088 Magnetic resonance imaging [MRI]
• CPC Class Codes: G06T 7/0012 Biomedical image inspection
• CPC Class Codes: G06T 7/12 Edge-based segmentation

Abstract

Structures are identified in pixel data representative of a discrete pixel image and are differentiated from non-structures. The structures are identified by comparing gradient magnitudes with a selected gradient value and by comparing gradient directions of adjacent pixels. Pixels having desired relationships are counted and a focus parameter is applied to the resulting count. A final gradient threshold value is identified based upon the count and the focus parameter. The threshold value identifies structural pixels. Isolated adjacent regions may be eliminated from the structures. Edges defined by the structures are binary rank order filtered to arrive at the final structural definition. The structural and non-structural pixels may be further processed to enhance the image for subsequent display.

Claims

1. A method for identifying structural image components in a discrete pixel image, the method comprising the steps of:

(a) computing gradient values for each pixel of the image, including gradient magnitude and direction values;

(b) identifying a gradient threshold value by selecting an initial threshold value prior to identifying the pixels representative of structural edges, counting candidate pixels representative of structural edges, and identifying a final gradient threshold value based upon the number of pixels representative of structural edges;

(c) based at least in part on the final gradient threshold value and the gradient magnitude and direction values, identifying pixels representative of structural edges; and

(d) binary rank order filtering the structural edges to identify pixels representative of structural components in the image.

2. The method of claim 1, wherein the final gradient threshold value is identified by adding a scalar value to the number of pixels representative of structural edges.

3. The method of claim 1, wherein the final gradient threshold value is identified by determining a gradient value equal to or less than gradient magnitude values for a desired number of pixels.

4. The method of claim 3, wherein the desired number of pixels is the number of pixels representative of structural edges plus a scalar value.

5. The method of claim 1, including the further step of eliminating from the structural components regions of adjacent pixels less than a desired count.

6. The method of claim 1, wherein the structural edges are identified by comparing gradient magnitude values for each pixel to a predetermined magnitude value, and comparing a gradient direction value for each pixel to gradient direction values of adjacent pixels.

7. A method for identifying pixels in a discrete pixel image representative of features of interest, the method comprising the steps of:

(a) determining gradient magnitude and direction values for each pixel;

(b) generating a gradient histogram for the pixels;

(c) identifying a gradient threshold value from the gradient histogram and from the gradient magnitude and direction values for each pixel; and

(d) identifying pixels representative of features of interest from the gradient threshold value and the gradient histogram.

8. The method of claim 7, wherein feature edges are identified by comparing gradient magnitude values for each pixel to a predetermined magnitude value, comparing a gradient direction value for each pixel to gradient direction values of adjacent pixels, and counting pixels satisfying predetermined relationships in the magnitude and direction comparisons.

9. The method of claim 8, wherein a final gradient threshold value is identified based upon the number of pixels representative of feature edges.

10. The method of claim 9, wherein the final gradient threshold value is identified by adding a scalar value to the number of pixels representative of feature edges.

11. The method of claim 10, wherein the final gradient threshold value is identified by determining a gradient value equal to or less than gradient magnitude values representing a desired number of pixels.

12. The method of claim 11, wherein the desired number of pixels is the number of pixels representative of feature edges plus a scalar value.

13. The method of claim 7, wherein the step of identifying pixels representative of features of interest includes the steps of identifying pixels representative of feature edges and binary rank order filtering the feature edges.

14. The method of claim 7, including the further step of eliminating from a group of pixels representative of features of interest regions of adjacent pixels less than a desired count.

15. A method for identifying pixels of interest in a discrete pixel image, the method comprising the steps of:

(a) determining gradient magnitude and direction values for each pixel;

(b) identifying contiguous pixels of interest based upon the gradient magnitude and direction values; and

(c) eliminating from the pixels of interest regions of adjacent pixels including a number of pixels less than a desired minimum number.

16. The method of claim 15, wherein the regions of adjacent pixels are identified by assigning index values to each pixel of interest, iteratively reassigning adjacent pixels a common index values, and counting the number of pixels sharing common index values.

17. The method of claim 16, wherein the common index values are assigned to adjacent pixels within an immediately adjacent neighborhood of the pixels of interest.

18. The method of claim 15, wherein the pixels of interest are identified by determining a gradient threshold and selecting pixels having a desired relationship to the gradient threshold.

19. The method of claim 18, wherein the gradient threshold is determined by identifying feature edges defined by the pixels, and counting the pixels defining the feature edges.

20. The method of claim 19, wherein the gradient threshold is determined by identifying a gradient value equal to or less than gradient magnitude values representing a desired number of pixels.

21. The method of claim 20, wherein the desired number of pixels is determined by counting the number of pixels defining the feature edges and adding a scalar value to the number.

22. An apparatus for identifying features of interest in a discrete pixel image, the apparatus comprising:

an image sensing device for generating raw data representative of a plurality of discrete pixels in an image matrix;

a signal processing circuit coupled to the sensing device, the signal processing circuit being configured to determine gradient magnitude and direction values for each pixel, generate a gradient histogram for the pixels, identify a gradient threshold value from the gradient histogram and from the gradient magnitude and direction values, and identify pixels representative of features of interest from the gradient threshold value and the gradient histogram; and

a memory circuit coupled to the signal processing circuit and configured to receive and save values representative of the pixels of interest.

23. The apparatus of claim 22, wherein the image sensing device includes a magnetic resonance scanner for detecting emissions of gyromagnetic materials in a subject of interest.

24. The apparatus of claim 22, further comprising an input device coupled to the signal processing circuit, the input device being configured to receive operator-defined parameters and to convey signals representative of the parameters to the signal processing circuit for identification of the pixels representative of features of interest.

25. The apparatus of claim 22, further comprising an output device coupled to the signal processing circuit, the output device being configured to generate an operator perceivable image reconstituted from data stored in the memory circuit.

26. An imaging system comprising:

a sensing station for detecting signals representative of discrete pixels in a subject of interest and for generating digital values representative of the detected signals;

a memory circuit for storing a predetermined routine for processing the digital values;

a signal processing circuit coupled to the sensing station and to the memory circuit, the signal processing circuit being configured to execute the routine to identify and enhance features of interest based upon a gradient threshold value determined from gradient magnitude and direction values for each digital value; and

an output device for producing an operator perceivable image including features of interest identified and enhanced by the signal processing circuit.

27. The imaging system of claim 26, wherein the sensing station includes a magnetic resonance scanner for detecting emissions of gyromagnetic material in the subject of interest at locations corresponding to the discrete pixels.

28. The imaging system of claim 26, wherein further comprising an input device coupled to the signal processing circuit for inputting parameters used by the signal processing circuit in identifying and enhancing the features of interest.