Detection of features in images

• US 7,519,206 B2
• Filed: 11/19/2001
• Issued: 04/14/2009
• Est. Priority Date: 11/22/2000
• Status: Expired due to Fees

• Current Assignee: Siemens Medical Solutions USA Incorporated (Siemens AG)
• Original Assignee: Siemens Medical Solutions USA Incorporated (Siemens AG)
• Inventors: Mulet-Parada, Miguel, Noble, Julia Alison
• Primary Examiner(s): Bella; Matthew C
• Assistant Examiner(s): Seth; Manav
• Application Number:
  US10/432,376
• Publication Number:
  US 20040047498A1
• Time in Patent Office: 2,703 Days
• Field of Search: 382128-132, 382/100, 382190-192, 382/203, 382260-265, 382/286, 600/407, 600437-444
• US Class Current: 382/128
• CPC Class Codes: G06T 2207/10132 Ultrasound image
• CPC Class Codes: G06T 2207/30048 Heart; Cardiac
• CPC Class Codes: G06T 7/0012 Biomedical image inspection
• CPC Class Codes: G06T 7/12 Edge-based segmentation
• CPC Class Codes: G06T 7/168 involving transform domain methods

Abstract

An image processing technique which identifies pixels in images which are associated with features having a selected shape, such as but not exclusively step edge, roof, ridge or valley. The shape of the intensity profile in the image is compared in an intensity independent way with a shape model to select those pixels which satisfy the shape model and are thus associated with the feature of interest. This comparison is achieved by examining the phase and amplitude of a spectral decomposition of parts of the image profile in the spatial or spatio temporal frequency domain. This decomposition can be achieved using quadrature wavelet pairs such as log-Gabor wavelets. The difference between the odd and even components, known as the feature asymmetry, gives an indication of the shape of the feature. The analysis may be extended to the time domain by looking at the shape of the image profile across a time sequence of images, which gives an indication of the velocity of a moving feature. Pixels identified as belonging to a feature of the right shape are labelled with the value of feature asymmetry, the local amplitude, feature orientation and feature velocity, and this information can be used to improve the tracking of detected features through a sequence of images.

Claims

1. A method of detecting features-of interest in an image, comprising:

using a processor to perform steps of;

processing the image to make an intensity independent representation of an image intensity profile;

comparing the intensity independent representation with an intensity independent shape model for predefined features of interest by;

convolving the intensity independent representation with a quadrature filter to derive the phase and amplitude of a pair of odd and even components of the profile in the spatial or spatio temporal frequency domains;

deriving from a difference between the components a measure of one of the feature symmetry or asymmetry along the profile,

deriving from the components a measure of the local amplitude along the profile; and

defining areas of the image satisfying a predefined constraint on the value of said one of the feature symmetry or asymmetry as satisfying the shape model and thus belonging to an image feature; and

storing for the profile a label comprising said measures.

2. A method according to claim 1, wherein the label further comprises a measure of feature orientation.

3. A method according to claim 1, wherein the predefined features of interest comprise positive and negative step edges in the intensity profile.

4. A method according to claim 1, further comprising performing the comparing on intensity independent representations derived across successive frames of an image sequence to derive a measure of a velocity of image features moving through the sequence, and wherein said label further comprises said measure of the velocity.

5. A method according to claim 1, wherein said quadrature filter is a log-Gabor filter.

6. A method according to claim 1, wherein said predefined constraint on said one of the feature symmetry or asymmetry is that its magnitude is greater than or equal to a predefined threshold.

7. A method according to claim 1, further comprising convolving the intensity independent representation with quadrature filters oriented in different directions across the image to detect image features oriented differently in the image, and deriving from relative responses of the filters a measure of feature orientation.

8. A method according to claim 1, further comprising tracking detected image features through successive frames of an image sequence by comparing labels of areas in a succeeding frame with labels of areas in the preceding frame over a predetermined search region to define as relating to the same feature those areas in each frame whose labels satisfy predefined search criteria.

9. A method according to claim 8, wherein the predetermined search region is defined as a region within a predetermined distance of a position of the detected image feature in an adjacent frame in the sequence.

10. A method according to claim 9, wherein the predetermined distance is set in dependence upon a velocity of the detected image feature.

11. A method according to claim 8, wherein the predefined search criteria comprise a condition on a value of feature symmetry or asymmetry of the areas in the search region.

12. A method according to claim 11, wherein the condition on the value of the feature symmetry or symmetry or asymmetry is that it is greater than a predetermined amount.

13. A method according to claim 11, wherein the predefined search criteria comprise a condition that a difference in orientation of the detected features in the two frames is less than a predetermined amount.

14. A method according to claim 1, wherein the comparing is performed at a plurality of scales.

15. A method according to claim 1, wherein the comparing is performed on an intensity profile taken through a plurality of images of different depths of a volume being imaged.

16. A method according to claim 1, wherein the image is an ultrasound or magnetic resonance or x-ray or standard camera image.

17. A method according to claim 1, wherein the image is an ultrasound contrast image, an ultrasound Doppler image or an ultrasound harmonic image.

18. A method according to claim 1, wherein the image is a medical image.

19. A method according to claim 1, wherein the image is a remote sensing or non-invasive image.

20. A method according to claim 1, wherein the image is an echocardiogram.

21. A method according to claim 19, wherein the feature is a ventricular or atrial wall.

23. An image processing system comprising a processor and a computer-readable medium storing a program which, when executed by the processor, performs the method of claim 1.

22. A computer program stored on a computer readable medium and comprising program code which is executable by a processor to perform a method of detecting features-of-interest in an image, the method comprising:

processing the image to make an intensity independent representation of an image intensity profile;

comparing the intensity independent representation with an intensity independent shape model for predefined features of interest by;

convolving the intensity independent representation with a quadrature filter to derive the phase and amplitude of a pair of odd and even components of the profile in the spatial or spatio temporal frequency domains;

deriving from a difference between the components a measure of one of the feature symmetry or asymmetry along the profile,

deriving from the components a measure of the local amplitude along the profile; and

defining areas of the image satisfying a predefined constraint on the value of said one of the feature symmetry or asymmetry as satisfying the shape model and thus belonging to an image feature; and

storing for the profile a label comprising said measures.