IDENTIFICATION OF REGIONS OF INTEREST AND EXTRACTION OF TIME VALUE CURVES IN IMAGING PROCEDURES

• US 20090316970A1
• Filed: 06/24/2008
• Published: 12/24/2009
• Est. Priority Date: 06/24/2008
• Status: Active Grant

• Current Assignee: Bayer Healthcare LLC (Bayer AG)
• Original Assignee: Medrad Incorporated (Bayer AG)
• Inventors: Kalafut, John F., Kemper, Corey
• Granted Patent: US 8,315,449 B2
• Time in Patent Office: Days
• US Class Current: 382/131
• CPC Class Codes: G06F 18/23213 with fixed number of clusters, e.g. K-means clustering
• CPC Class Codes: G06T 2207/10081 Computed x-ray tomography [CT]
• CPC Class Codes: G06T 2207/30061 Lung
• CPC Class Codes: G06T 7/0012 Biomedical image inspection
• CPC Class Codes: G06T 7/0016 involving temporal comparison
• CPC Class Codes: G06T 7/11 Region-based segmentation
• CPC Class Codes: G06T 7/143 involving probabilistic approaches, e.g. Markov random field [MRF] modelling
• CPC Class Codes: G06V 10/763 Non-hierarchical techniques, e.g. based on statistics of modelling distributions
• CPC Class Codes: G06V 2201/031 of internal organs

Abstract

A method of extracting at least one time-value curve to determine a protocol in an imaging procedure using an imaging system, includes: determining a first N-dimensional data set of pixel values of a portion of a body of the patient at a first time using the imaging system, wherein N is an integer; determining at least a second N-dimensional data set of pixel values of the portion at a second time using the imaging system; computing a predetermined number of correlated segments of the imaged portion corresponding to a predetermined number of regions of interest of the patient by computing a similarity metric of a time series of pixel values; computing the at least one time-value curve for at least one of the regions of interest; and determining a protocol for a diagnostic scan using the image system based at least in part upon data from the time value curve.

Claims

1. A method of extracting at least one time-value curve to determine a protocol in an imaging procedure using an imaging system, comprising:

determining a first M-dimensional data set of pixel values of a portion of a body of the patient at a first time using the imaging system, wherein M is an integer;

determining at least a second M-dimensional data set of pixel values of the portion at a second time using the imaging system;

computing a predetermined number of correlated segments of the imaged portion corresponding to a predetermined number of regions of interest of the patient by computing a similarity metric of a time series of pixel values;

computing the at least one time-value curve for at least one of the regions of interest; and

determining a protocol for a diagnostic scan using the image system based at least in part upon data from the time value curve.

2. The method of claim 1 wherein a series of 0 through T M-dimensional datasets are determined, wherein M is an integer and T is an integer, and wherein the 0 dataset corresponds to a dataset at time t=0 and the T dataset corresponds to a dataset at time t=T.

3. The method of claim 2 wherein a vector y_x=[y_x(0) y_x(1) . . . y_x(T−

1) y_x(T)] is determined wherein y is a value of a pixel at location x.

4. The method of claim 3 wherein the segments are computed using K-means clustering.

5. The method of claim 4 wherein K clusters are computed by minimizing the sum of squares of distances between data points and K cluster centroids, wherein K is an integer corresponding to the predetermined number of correlated segments.

6. The method of claim 5 wherein pixel values are randomly assigned to one of K clusters and each cluster centroid is a component-wise mean of the pixel values in that cluster after centering and normalizing pixel values to a zero mean and unit standard deviation as follows:

7. The method of claim 6 wherein each cluster centroid is determined using the following formula:

8. The method of claim 7 wherein the distance from centered and normalized pixel values to each of the K cluster centroids Y_c is determined using the formula:

9. The method of claim 8 wherein each centered and normalize pixel value is assigned to the cluster to which it exhibits the minimum computed distance to the centroid.

10. The method of claim 9 wherein the actions of claims 7 through 9 are repeated until convergence, resulting in a segmented image having K segments corresponding to the determined K clusters.

11. The method of claim 10 further comprising filtering the segmented image to eliminate pixels that are not well correlated with neighboring pixels.

12. The method of claim 11 wherein each pixel of the segmented image is compared with all eight of its neighboring pixels in filtering.

13. The method of claim 11 further comprising morphologically opening the segmented image.

14. The method of claim 13 further comprising overlaying the segmented image upon a contrast enhanced bolus data set and computing enhancement profiles for each of the K segments.

15. The method of claim 14 further comprising semantically labeling at least one of the enhancement profiles with a semantic label corresponding to a region of interest of the patient based upon at least one of the characteristics of the labeled enhancement profile as compared to at least one other computed enhancement profiles.

16. The method of claim 1 further comprising overlaying the segmented image upon a contrast enhanced bolus data set and computing enhancement profile for each of the K segments.

17. The method of claim 16 further comprising the step of semantically labeling at least one of the enhancement profiles with a semantic label corresponding to a region of interest of the patient based upon at least one of the characteristics of the labeled enhancement profile as compared to at least one other computed enhancement profiles.

18. A system for extracting at least one time-value curve to determine a protocol in an imaging procedure, comprising:

an input system for input of data output from at least one imaging system, the data comprising a first M-dimensional data set of pixel values of a portion of a body of the patient at a first time, wherein M is an integer, and at least a second M-dimensional data set of pixel values of the portion at a second time;

at least one processor in communicative connection with the input system and adapted to compute a predetermined number of correlated segments of the imaged portion corresponding to a predetermined number of regions of interest of the patient by computing a similarity metric of a time series of pixel values and computing the at least one time value curve for at least one of the regions of interest; and

at least one parameter generator system to determine a protocol for a diagnostic scan using the image system based at least in part upon data from the time value curve.

19. The system of claim 18 further comprising a memory system in communicative connection with the at least one processor, the memory system having stored therein an algorithm executable by the at least one processor to compute the predetermined number of correlated segments of the imaged portion corresponding to the predetermined number of regions of interest of the patient.

20. The system of claim 18 wherein the system is in communicative connection with the at least one imaging system.

21. The system of claim 18 wherein the system is in communicative connection with the at least one injector system.

22. The system of claim 18 wherein at least a portion of the system is integrated with the at least one imaging system.

23. The system of claim 18 wherein at least a portion of the system is integrated with the at least one injector system.