Segmentation of lesions in ultrasound images

• US 20060247525A1
• Filed: 04/28/2006
• Published: 11/02/2006
• Est. Priority Date: 04/28/2005
• Status: Active Grant

• Current Assignee: CareStream Health Incorporated
• Original Assignee: CareStream Health Incorporated
• Inventors: Liu, Xu, Huo, Zhimin
• Granted Patent: US 8,517,945 B2
• Time in Patent Office: Days
• US Class Current: 600/437
• CPC Class Codes: G06T 2207/10132 Ultrasound image
• CPC Class Codes: G06T 2207/20036 Morphological image processing
• CPC Class Codes: G06T 2207/30068 Mammography; Breast
• CPC Class Codes: G06T 7/0012 Biomedical image inspection
• CPC Class Codes: G06T 7/11 Region-based segmentation
• CPC Class Codes: G06T 7/162 involving graph-based methods

Abstract

A method for determining a candidate lesion region in a digital ultrasound medical image of anatomical tissue. The method includes the steps of: accessing the digital ultrasound medical image of anatomical tissue; applying an anisotropic diffusion filter to the ultrasound image to generate a filtered ultrasound image; performing a normalized cut operation on the filtered ultrasound image to partition the filtered ultrasound image into a plurality of regions; and selecting, from the plurality of regions, at least one region as a candidate lesion region.

Claims

1. A method for determining a candidate lesion region in a digital ultrasound medical image of anatomical tissue, the method comprising the steps of:

accessing the digital ultrasound medical image of anatomical tissue;

applying an anisotropic diffusion filter to the ultrasound image to generate a filtered ultrasound image;

performing a normalized cut operation on the filtered ultrasound image to partition the filtered ultrasound image into a plurality of regions; and

selecting, from the plurality of regions, at least one region as a candidate lesion region.

2. The method of claim 1, wherein the selected candidate lesion region has:

(1) an intensity value lower than a pre-determined intensity value and (2) morphological or texture features in accordance with pre-determined lesion criteria.

3. The method of claim 1, further comprising the step of, prior to the selecting step, merging the plurality of regions based on pre-determined threshold values.

4. The method of claim 3, wherein a first threshold value is applied to the plurality of regions having an average gray value lower than a pre-determined value G₀, and a second threshold value is applied to the plurality of regions having an average gray value higher than pre-determined value G₀.

5. The method of claim 1, further comprising the step of segmenting the selected at least one candidate lesion region in the ultrasound image.

6. The method of claim 1, wherein the step of performing a normalized cut is accomplished by segmenting spatially contiguous pixels in the filtered ultrasound image into a plurality of regions in accordance with substantially similar features.

7. The method of claim 6, wherein the normalized cut is accomplished to minimize a similarity between each of the plurality of regions and maximize a similarity within each of the plurality of regions.

8. The method of claim 7, wherein the features include intensity-based features and/or texture-based features.

9. A method for determining a candidate lesion region in a digital ultrasound medical image of anatomical tissue, the method comprising the steps of:

accessing the digital ultrasound medical image of anatomical tissue;

applying an anisotropic diffusion filter to the ultrasound image to generate a filtered ultrasound image;

performing a normalized cut operation on the filtered ultrasound image to partition the filtered ultrasound image into a plurality of regions, wherein the normalized cut is performed by segmenting spatially contiguous pixels in the filtered ultrasound image into a plurality of regions in accordance with substantially similar features;

merging the plurality of regions based on pre-determined threshold values; and

selecting, from the plurality of regions, at least one region as a candidate lesion region, wherein the selected candidate lesion region has;

(1) an intensity value lower than a pre-determined intensity value and (2) morphological or texture features in accordance with pre-determined lesion criteria.

10. The method of claim 9, wherein a first threshold value is applied to the plurality of regions having an average gray value lower than a pre-determined value G₀, and a second threshold value is applied to the plurality of regions having an average gray value higher than pre-determined value G₀.

11. The method of claim 9, further comprising the step of segmenting the selected at least one candidate lesion region in the ultrasound image.

12. The method of claim 9, wherein the normalized cut is accomplished to minimize a similarity between each of the plurality of regions and maximize a similarity within each of the plurality of regions.

13. The method of claim 12, wherein the features include intensity-based features and/or texture-based features.