Shape based similarity of continuous wave doppler images

• US 8,712,122 B2
• Filed: 03/31/2011
• Issued: 04/29/2014
• Est. Priority Date: 03/31/2011
• Status: Active Grant

• Current Assignee: International Business Machines Corporation
• Original Assignee: International Business Machines Corporation
• Inventors: Beymer, David James, Syeda-Mahmood, Tanveer F.
• Primary Examiner(s): Fitzpatrick, Atiba O
• Assistant Examiner(s): JAMES, THOMAS AIDEN
• Application Number:
  US13/076,981
• Publication Number:
  US 20120250957A1
• Time in Patent Office: 1,125 Days
• Field of Search: None
• US Class Current: 382/128
• CPC Class Codes: A61B 8/0883 for diagnosis of the heart
• CPC Class Codes: A61B 8/488 involving Doppler signals
• CPC Class Codes: A61B 8/5223 for extracting a diagnostic or physiological parameter from medical diagnostic data for algorithms to analyse biomedical images G06T7/0012
• CPC Class Codes: G06T 2207/10132 Ultrasound image
• CPC Class Codes: G06T 2207/30048 Heart; Cardiac
• CPC Class Codes: G06T 7/0016 involving temporal comparison
• CPC Class Codes: G06V 2201/03 Recognition of patterns in medical or anatomical images

Abstract

Continuous wave Doppler images in a data base comprising cardiac echo studies are processed to separate Doppler frames. The frames are pre-processed to extract envelope curves and their corner shape features. Shape patterns in Doppler images from echo studies of patients with known cardiac (valvular) diseases are employed to infer the similarity in valvular disease labels for purposes of automated clinical decision support. Specifically, similarity in appearance of Doppler images from the same disease class is modeled as a constrained non-rigid translation transform of velocity envelopes embedded in these images. Shape similarity between two Doppler images is then judged by recovering the alignment transform using a variant of dynamic shape warping. Since different diseases appear as characteristic shape patterns in Doppler images, measuring the similarity in the shape pattern conveyed within the velocity region of two Doppler images can infer the similarity in their diagnosis labels.

Claims

1. A method comprising:

storing in a database at least one continuous wave (CW) Doppler image and at least one characteristic associated with each of the at least one stored Doppler image in a form of an envelope curve;

gathering, by a processor, a query Doppler image, and extracting the envelope curve from a velocity profile associated with a query image;

matching, by the processor, the at least one characteristic associated with said at least one stored image with at least one characteristic associated with the envelope curve extracted from the query image; and

retrieving, by the processor, from the database at least one image having characteristics matching the characteristics associated with the query image.

2. The method of claim 1, further comprising:

extracting said (CW) Doppler image from an echo-cardiogram video; and

isolating the velocity profile from said extracted Doppler image.

3. The method of claim 1, further comprising determining a heart beat cycle within said envelope curve of the query image and normalizing the envelope curve to include at least one heart beat cycle and a fixed image size.

4. The method of claim 3, further comprising each characteristic associated with the envelope curve reflecting a shape of the envelope curve, wherein matching characteristics associated with a stored image and the characteristics associated with the query image is based upon a shape similarity of the envelope curve of the query image and an envelope curve of the stored image.

5. The method of claim 1, further comprising determining a shape similarity of the envelope curve of the at least one stored image with the query image by selecting a non-rigid time transform to minimize a difference between the shapes of the envelope curve of the query image and the envelope curve of the at least one stored image.

6. The method of claim 5, further comprising representing said at least one characteristic associated with said envelope curves of the query and stored images through fiducial points, said fiducial points chosen as corners in a line segment approximation for each envelope curve.

7. A system comprising:

a database to store at least one continuous wave (CW) Doppler image and at least one characteristic associated with each of the at least one stored Doppler image in a form of an envelope curve;

a query manager to gather a query Doppler image and to extract the envelope curve from a velocity profile associated with the query image;

an image manager to match the at least one characteristic associated with the stored image with at least one characteristic associated with the envelope curve extracted by the query manager from the query image; and

a retrieval manager in communication with the image manager, the retrieval manager to retrieve from the database at least one image having a characteristic matching the characteristic associated with the query image.

8. The system of claim 7, further comprising:

the query manager to extract said Doppler image from an echo-cardiogram video; and

the query manager to isolate the velocity profile from said extracted Doppler image.

9. The system of claim 7, further comprising the query manager to determine a heart beat cycle within said envelope curve of the query image and to normalize the envelope curve to include at least one heart beat cycle and a fixed image size.

10. The system of claim 7, further comprising each characteristic associated with the envelope curve to reflect a shape of the envelope curve, wherein the image manager match of the characteristic associated with a stored image and the characteristic associated with the query image is based upon a shape similarity of the envelope curve of the query image and an envelope curve of the at least one stored image.

11. The system of claim 7, further comprising the image manager to determine a shape similarity of the envelope curve of the stored image with the query image by selection of a non rigid transform to minimize a difference between the shapes of the envelope curve of the query image and the envelope curve of the stored image.

12. The system of claim 11, further comprising the image manager to represent said at least one characteristic associated with said envelope curves of the query and stored images through fiducial points, said fiducial points chosen as corners in a line segment approximation for each envelope curve.

13. An article comprising:

a non-transitory computer readable data storage media, said media including instructions comprising;

instructions for storing in a database at least one continuous wave (CW) Doppler image and at least one characteristic associated with each of the at least one stored Doppler image in a form of an envelope curve;

instructions for gathering, by a processor, a query Doppler image, and extracting the envelope curve from a velocity profile associated with the query image;

instructions for matching, by the processor, the at least one characteristic associated with the stored image with at least one characteristic associated with the envelope curve extracted from the query image; and

instructions for retrieving, by the processor, from the database at least one image having characteristics matching the characteristic associated with the query image.

14. The article of claim 13, further comprising:

instructions for extracting said Doppler image from an echo-cardiogram video; and

instructions for isolating the velocity profile from said extracted Doppler image.

15. The article of claim 13, further comprising instructions for determining a heart beat cycle within said envelope curve of the query image and normalizing the envelope curve to include at least one heart beat cycle and a fixed image size.

16. The article of claim 13, further comprising each characteristic associated with the envelope curve reflecting a shape of the envelope curve, wherein matching the at least one characteristic associated with a stored image and the at least one characteristics associated with the query image is based upon a shape similarity of the envelope curve of the query image and an envelope curve of the at least one stored image.

17. The article of claim 13, further comprising instructions for determining a shape similarity of the envelope curve of the stored image with the query image by selecting a non rigid transform to minimize a difference between the shapes of the envelope curve of the query image and the envelope curve of the stored image.

18. The article of claim 17, further comprising instructions for representing said at least one characteristic associated with said envelope curves of the query and stored images through fiducial points, said fiducial points chosen as corners in a line segment approximation for each envelope curve.