WO1998024363A1 - Method and apparatus for cumulative distribution image boundary detection - Google Patents

Method and apparatus for cumulative distribution image boundary detection

Info

Publication number
WO1998024363A1
WO1998024363A1 PCT/US1997/004871 US9704871W WO1998024363A1 WO 1998024363 A1 WO1998024363 A1 WO 1998024363A1 US 9704871 W US9704871 W US 9704871W WO 1998024363 A1 WO1998024363 A1 WO 1998024363A1
Authority
WO
Grant status
Application
Patent type
Prior art keywords
power
value
blood
reference
region
Prior art date
Application number
PCT/US1997/004871
Other languages
French (fr)
Inventor
Jonathan M. Rubin
Ray Steven Spratt
J. Brian Fowlkes
Ronald S. Adler
Original Assignee
The Regents Of The University Of Michigan
Priority date (The priority date is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the date listed.)
Filing date
Publication date

Links

Classifications

    • AHUMAN NECESSITIES
    • A61MEDICAL OR VETERINARY SCIENCE; HYGIENE
    • A61BDIAGNOSIS; SURGERY; IDENTIFICATION
    • A61B8/00Diagnosis using ultrasonic, sonic or infrasonic waves
    • A61B8/06Measuring blood flow

Abstract

An automated method for determining the boundary between a region of blood and a region of tissue helps overcome the shortcomings of conventional methods. The method entails finding a reference power value between two separate regions within the same image, or data set. The reference power value is selected in an automated fashion from a region within a region of 100 percent blood flow and a region of tissue. A cumulative power distribution curve is generated (420) from Doppler power values corresponding to the regions of 100 percent blood flow and tissue. The cumulative power distribution curve is used to determine (440) the reference power value which indicates the boundary between the two regions. The reference power value may be used to set a green tag or to numerically indicate values that correspond to blood. If the values corresponding to blood are identified, a second cumulative power distribution curve may be generated enabling a more stable reference power value to be determined. The second reference power value may be used to estimate the power value at a depth given rouleaux formation.

Description

METHOD AND APPARATUS FOR CUMULATIVE DISTRIBUTION IMAGE BOUNDARY DETECTION

1. BACKGROUND OF THE INVENTION This application claims priority of an application filed in the United States (U.S. Serial

Number 08/760,917) on December 6, 1996, which is hereby incorporated by reference in its entirety.

The invention relates in general to the field of medicine, and more particularly, to boundary detection in medical imaging of regions of organisms containing both tissues and blood flow. Specifically, the invention relates to an automated method for determining a value that represents 100 percent blood flow and to an apparatus for performing the method.

In the diagnosis of various medical conditions, it is often useful to examine soft tissue and blood flow within the body to show structural details of organs and the corresponding blood vessels. Multiple studies have demonstrated increased vascularity (i.e., blood flow) in many tumors relative to that of normal tissue. Numerous attempts have been made to depict these differences in vascularity using ultrasonic imaging. Typically, techniques for quantifying blood flow (e.g., fractional moving blood volume estimation) use a value representing 100 percent blood flow to normalize measurements. A commonly-assigned pending application serial number 08/657,897, describing fractional moving blood volume estimation, is hereby incorporated by reference in its entirety.

Normalization generally involves making a measurement of the blood flow and dividing that number by a reference blood value representing 100 percent blood flow at a given depth. Due to the attenuation of imaging signals, normalization generally is done at each depth of interest. Normalization enables the comparison of vascularity in two different types of tissues. For example, the vascularity of kidney may be compared to the vascularity of the liver. Even if they were located at different depths, normalization would enable comparison that could indicate the vascularity of one with respect to the other.

Often it can be difficult to determine a reference blood value for 100 percent blood flow. One of the reasons is because of rouleaux formation in blood, which is a loose adhesion of red blood cells that form long chains as illustrated in Figure 1. The adhesion of red blood cells can make finding a reference blood value more difficult because blood vessels produce a distribution of values. Conventional prior-art methods for selecting a reference blood value representing 100 percent blood flow generally require operators to visually select a region that they consider to be 100 percent and use the power value associated with that region as the reference blood value. Reference blood values based on visual selection are more prone to contain error which reduces accuracy in the measurement.

Conventional selection methods have also depended on the location of the reference vessel. Though some methods require that the reference vessel be a part of the region of interest, that may not always be optimal. For example, a reference vessel within the region of interest may give a less accurate measurement than a larger vessel located outside the region of interest. If a method only allows the use of reference vessels located within the region of interest, the method may be prone to more restrictive use in regions with large blood vessels.

The determination of a reference blood value generally occurs after the boundary between a region of tissue and a region of blood has been found. Thus, conventional boundary tracking methods (e.g., gradient techniques) could also be used to determine a reference blood value which could indicate the boundary between two regions. However, boundary tracking methods are more susceptible to error, because they are noise sensitive and often coordinate system dependent. As a result, they can suffer from local variations.

2. SUMMARY OF THE INVENTION The invention relates to an automated method for determining the boundary between a region of blood and a region of tissue that helps to overcome the shortcomings of conventional methods. The method entails finding a reference power value between two separate regions within an image (i.e., a data set). The reference power value is selected in an automated fashion from a region with a region of 100 percent blood flow and a region of tissue. A cumulative power distribution curve is generated from Doppler power values corresponding to the regions of 100 percent blood flow and tissue. The cumulative power distribution curve is used to determine the reference power value which indicates the boundary between the two regions.

The reference power value may be used to set a green tag or to numerically indicate values that correspond to blood (as opposed to tissue). If the values corresponding to blood are identified, a second cumulative power distribution curve may be generated enabling a more stable reference power value to be determined. The second reference power value may be used to estimate the power value in a vessel to compensate for rouleaux formation. 3. BRIEF DESCRIPTION OF DRAWINGS

Figure 1 illustrates rouleaux formation in a blood vessel.

Figures 2 and 3 illustrate components of an imaging system. 5 Figure 4 is a flow chart illustrating steps in a first method in accordance with the invention.

Figures 5 through 9 illustrate detailed views of each of the steps in the first method.

Figure 10 is a flow chart illustrating steps in a second method in accordance with the invention.

Figure 11 illustrates a detailed view of the last step in the second method.

Figure 12 is a flow chart of the steps for a third embodiment of the invention. l o 4. DETAILED DESCRIPTION OF SPECIFIC EMBODIMENTS

Illustrative embodiments of the invention are described below as they might be employed in determining a boundary between a region of 100 percent blood flow and a region of tissue. In the interest of conciseness, not all features of an actual implementation are described in this specification. It will of course be appreciated that in the development of any such actual 15 embodiment numerous implementation-specific decisions must be made to achieve the developers' specific goals, such as compliance with system-related and business-related constraints. Moreover, it will be appreciated that even if such a development effort might be complex and time-consuming it would nevertheless be a routine undertaking for those of ordinary skill having the benefit of this disclosure.

20 4.1 Overview

The claimed invention relates to finding the boundary between a region of blood and a region of tissue in a scan of a patient such as an ultrasound scan. Finding that boundary is useful in determining which portion of the ultrasound scan signals are from regions of 100 percent blood flow (e.g., arteries and veins) and which are from regions that are tissue (e.g., liver). The

25 invention helps overcome some of the shortcomings of conventional methods of finding the boundary.

In one implementation of the invention, a scanning subsystem 205 is used to scan a patient 300 (see Figures 2, 3). An image 330, generated from the scan data, is displayed on a screen 220. The operator of the scanning subsystem electronically marks off a region of interest

30 (ROI) 510 in the image 330 (see Figure 5). From then on, the processing is automatically performed by a processor 210 in an ultrasound machine 320 under the control of special programming in the program storage device 215. The processing system 210 plots a cumulative power distribution curve 600 for the region of interest 510 (see Figure 6). It analyzes the "knee" of the curve to locate the power level that corresponds to the desired blood-tissue boundary .

One way in which the processing system 210 can analyze the knee of the curve 600 is to mathematically fit a straight line 700 to the curve (see Figure 7), and to rotate the curve so that the straight line lies on the horizontal axis of the plot, which is referred to as the abscissa axis of the plot (see Figure 8). The machine 320 then mathematically fits a parabola to the rotated curve (see Figure 9). It selects the power level corresponding to the top of the parabola 920 (shown as power level br in Figure 9) as the power level of the blood/tissue boundary. A second way in which the processing system 210 can analyze the knee of the curve 600 is to mathematically fit straight tangent lines T and Tc2 to the two different "limbs" of the curve 630, 640 (see Figure 1 1). The processing system selects the power level corresponding to the intersection point 1 105 of the two tangent lines (shown as power level b in Figure 1 1) as the power level of the blood tissue boundary. If either the first or the second way is used, the basic process can be repeated for the blood portion of the data to obtain a better estimate of the "normalization value."

4.2 Apparatus

The primary components of an imaging system 200 are shown in Figure 2. A scanning subsystem 205 may be used to perform a scan of the media of interest to acquire data to be used in the generation of an image of the media. The scanning subsystem 205 can vary in size depending on the actual type of scan being performed. For example, MRI (i.e., magnetic resonance imaging) scans are performed by having the patient (i.e., the media) lay down on a table in which case the scanning subsystem is normally very large. In contrast, the scanning subsystem used in ultrasonic imaging is generally smaller, often referred to as a scanhead. The programmable processing system 210 can process the data received from the scanning subsystem. This system may be either internal or external to the imaging system 200. For example, an ultrasonic imaging system may have an internal processor to process the data or it may be attached to a separate computer that has a processor.

Programming of the processing system 210 may be accomplished through the use of a program storage device 215 readable by the processor and encoding a program of instructions executable by the processor for performing the operations described below. The program storage device may take the form of, e.g., one or more floppy disks; a CD ROM or other optical disk; a magnetic tape; a read-only memory chip (ROM); and other forms of the kind well-known in the art or subsequently developed. The program of instructions may be "object code," i.e., in binary form that is executable more-or-less directly by the processor; in "source code" that requires compilation or interpretation before execution; or in some intermediate form such as partially compiled code. The precise forms of the program storage device and of the encoding of instructions are immaterial here.

An advantage of the invention is that many conventional imaging systems 200 can be upgraded by reprogramming their processor systems 210 to perform the operations described here. An imaging system arranged as shown in Figure 2 may be beneficial when it is desirable to obtain and record scan data. This data can be processed and analyzed "off-line" (i.e., external to the imaging system) either at the same location or a different location. The data can be transferred via an on-line connection or via a storage device such as a disk or magnetic tape.

The imaging system 200 shown in Figure 4 may be used by an operator (not shown), who holds a transducer 205 in one position relative to a volume of material 300 (e.g., human tissue). The transducer 205 is sometimes referred to as a scanhead; it commonly has an essentially linear, one-dimensional (ID) shape, although scanheads of round or other shapes are also known, and emit a beam of ultrasound energy toward the material 300. The ultrasound energy is reflected from various portions of the material 300 and detected by the scanhead 205, which generates data signals representative of the detected energy. A conventional ultrasound machine 320 receives and processes the resulting data from the scanhead 205 and displays a 2D image 330 of the tissue volume 300 being scanned (e.g., on a video display terminal 220), a film camera, or other hard copy device (not shown). Movement of the scanhead 205 results in different 2D images 330 of the tissue volume 300 being presented.

4.3 Fitting a Parabola to the Rotated Curve

Figure 4 is a flow chart providing a high-level illustration of one method in accordance with the invention. The operations performed in the method include conventionally generating an image at block 410, selecting an ROI (region of interest) at block 415, generating a cumulative power distribution curve at block 420, fitting a line at block 430, rotating the distribution at block 435, and determining a reference power value at block 440.

The image 330 generated at block 410 may vary in power (i.e., intensity) levels and contrast depending on the type of imaging technique selected. As will be apparent to those of ordinary skill in the art having the benefit of this disclosure, other types of imaging (e.g., NMR imaging and x-ray imaging) may also be used instead of ultrasound imaging to generate images. Multiple rows and columns are often use to divide the screen of a display into subelements, or pixels, that are controlled by the processing system.. After an image 330 has been generated, an ROI (region of interest) can be selected within a given image. The operator of an imaging machine can select the ROI based on a particular parameter of interest. The ROI 510 may be manually selected by an operator using a pointing device (e.g., mouse or track ball). The selected ROI 510 contains vessels with both 100 percent blood flow 520 and tissue 530. The power level of each individual pixel within the ROI may be stored in a storage device such as a hard disk (not shown). Computer software executing on the processing system 210 (e.g., Mathlab with suitable configuration) may be used to generate a cumulative power distribution curve 600 from the Doppler power values as shown in Figure 6. The abscissa (i.e., horizontal) axis 610 has units of Doppler power values and the ordinate (i.e., vertical) axis 620 has units of pixels .

A point on the cumulative power distribution curve indicates the number of pixels with power values less than or equal to a given power value as shown in Figure 6. The point A] on the cumulative power distribution curve 600 indicates that there are Yi pixels with power values less than or equal to Xi as shown by the shading. (In contrast, on a "regular" power curve point At might indicate that there are Yt pixels with power values equal to Xi).

The cumulative power distribution curve 600 normally will be composed of two separate regions as illustrated in Figure 6. The first region 630, with power values less than some as-yet unknown value b, corresponds to the tissue region. Values greater than the unknown value b indicate the second region 640 corresponding to the 100 percent blood region. In other words, the intersection between the portion of the cumulative power distribution curve associated with the tissue region and the portion of the cumulative power distribution curve associated with the blood region defines a boundary. Anything to the right of this boundary is essentially all blood and to the left is essentially all tissue; the abscissa value (shown in the Figure 6 as b) of the boundary is used as a reference power value. Referring to Figure 7, once the cumulative power distribution curve plot has been generated, the curve can be fit (block 330) with a line 700 using the method of least squares, also known as linear regression. As is well known in the art, linear regression involves minimizing the error of estimation (i.e., the difference between the actual value and the estimated value). As illustrated in Figure 7, the method of least squares involves minimizing the difference between the estimated function (i.e., the line) 700 and the actual function (i.e., the curve) 600.

The line 700 intersects the distribution is intersected at two points Cj and C2. These points are located in their associated regions (i.e., Cj is in the tissue region 630 and C2 is in the blood region 640).

Though the suspected location of the boundary is within the portion of the curve 600 between Cl and C2, the specific location of the reference power value b is not known. As noted above, the reference power value b is the power value above which is essentially all blood and below which is essentially all tissue. Rotating the distribution (i.e., block 335) is performed by using basic geometry and trigonometry to determine the angle (not shown) that the fitting line 600 makes with the abscissa axis. Once the angle is found, the entire distribution can be rotated using conventional techniques until C[ and C2 are on the abscissa axis as shown in Figure 8. The rotated cumulative power distribution curve 800 enables more accurate assessment of the specific location of the reference power value.

After the cumulative power distribution curve is rotated, the boundary point can be determined (block 340). As shown in Figure 9, the region 900 surrounding the highest point of the rotated distribution 800 is fitted with a parabola 910 using the method of least squares. Basic algebra is used to determine the equation of the parabola. The power level of the maxima (i.e., highest point) 920 of the parabola is used as the power level br boundary point; it indicates the junction between the separate regions (i.e., the blood region 630 and the tissue region 640). The power level of the maxima 920 may also be visually determined. This abscissa value (shown as bβ) of the maxima 920 is used as the reference power value.

The reference power value may be used to quantitatively identify the tissue region from the blood region because it is the power level at the junction between the tissue and blood region; thus, power levels above represent essentially all blood and below represent essentially all tissue. The reference power value may also be used to define the reference blood value (i.e., the best estimate of the power value for blood flow at a specific depth given rouleaux formation). When a reference blood value is defined, a "green tag" may be set for depth normalization.

4.4 Fitting Tangent Lines to the Curve

Figure 10 illustrates a second embodiment of the invention. The operations in a method 1000 of determining a boundary include: generating an image at block 410, selecting an ROI (region of interest) at block 415, generating a cumulative power distribution curve at block 420, and fitting a line at block 430. Tangent lines are constructed at block 1010 and a reference power value is determined at block 1040. The first four operations 410 through 430 for this method are identical to the first embodiment. Symmetric points surrounding each point of intersection are used to determine local tangent lines to intersection points Cj and C2 as shown in Figure 10. The local tangent lines TC1 and TC2 extend until they intersect at the boundary point 1105 between the tissue region and the blood region. As before, the reference power value is determined in step 1040 from the abscissa value (shown as b) of the boundary. In one preferred embodiment, tangent lines are calculated based on the best-fit line to a cumulative power distribution curve. The tangent lines are determined by locating the points of intersection of the global best- fit line with the distribution. For each sequential power value, the corresponding value in the cumulative power distribution curve is identified using the distribution itself and also the calculated estimate using the best-fit line. Then for each power value, the difference between the true distribution measurement and the estimate from the best-fit line is taken; any change of sign of this difference is noted. The points at which the sign of the difference changes are defined as the points of intersection of the best-fit line and the distribution.

P'(x) = Ax +B is the form of the best-fit line for a given distribution where P'(x) is the estimated cumulative value for a given x (i.e., Doppler power value). If P(x) is the true distribution from the data, then P(x) - P'(x) will be zero at points of intersection of the true distribution and the best-fit line. Since real data makes it difficult to rely on the actual points of intersection being in the data set, it is beneficial to identify the closest point in the data set to the intersection as being the one where there is a sign change. For example, an intersection would lie immediately between points x(i) and x(i+l) if (P(x(i) ) - P'(x(i) )) < 0 and (P(x(i+1)) - P'(x(i+1))) > 0. As a first approximation which is accurate enough for the current use of the intersection, either x(i) or x(i+l) could be chosen as the intersection. After selection of either one of these points at each intersection, a 21 -point region of interest centered about each selected point (e.g., x(i) + 10 points) is selected; each group of 21 points is fitted with a line. These local fitted lines are the local tangent.

There could be multiple intersections of the global best-fit line with the cumulative power distribution curve. To minimize the likelihood of this, each point of intersection of the best-fit line with the cumulative power distribution curve moving from the origin toward increasing abscissa values (i.e., Doppler power values) is identified. The first pair of lines in which the difference between the slopes is negative (i.e., the slopes are decreasing) is selected. For example, if T is a tangent line and i = {set of abscissa values of intersection points ranging sequentially from 1 to n}, then T(i) is the tangent line to the cumulative power distribution curve based around i. If Slope T(i) is the slope of tangent line T(i) and if T(i) and T(i+1) are the first pair of tangent lines encountered moving towards the right from the origin for whose slopes (Slope T(i) - Slope T(i+1)) < 0, then the intersection of T(i) and T(i+1) is the chosen point for the knee of the cumulative power distribution curve.

4.5 Two-Stage Analysis

Figure 12 illustrates a third method in accordance with the invention. Once a line 700 has been fitted to the cumulative power distribution curve, an initial reference power value is determined using either the parabola-fitting approach or the tangent-fitting approach described above. Once an initial reference power value has been determined, blood vessels are distinguished from surrounding tissue by discarding power values less than the reference power value at block 1205.

A second cumulative power distribution curve is then generated from the same data, but this time using only the power-value data corresponding to the blood vessels as opposed to surrounding tissue. That curve is linearly fitted using the method of least squares, and tangent lines are constructed to determine a second reference power value in the same general manner as described above. (If the data points in the curve are too few for an acceptable tangent-line fitting, the power values of those data points may simply be averaged and the resulting average value defined as the second reference power value.) The second reference power value defines a reference blood value at block 1220. It should be noted that the separation of the Doppler power values of surrounding tissue from the power values of blood vessels compensates at least in part for rouleaux formation, thus yielding a more stable reference blood value. When a reference blood value is determined, a "green tag" can be set.

4.6 Remarks Increased media coverage has led to a greater focus on vascularity estimates of ultrasound. Fractional moving blood volume estimation (FMBVE) is a technique that estimates vascularity using Power Doppler ultrasound while overcoming inherent limitations of other conventional estimating techniques. Proper use of FMBVE requires the determination of a reference blood value corresponding to 100 percent blood for depth normalization.

An automated method, in accordance with the invention, provides a stable reference blood value while accurately estimating vascularity. Moreover, a device in accordance with the invention may also be used to accurately determine the boundary between two different regions in which conventional boundary tracking methods would produce less than optimal results. The stability of the reference blood value resulting from the present invention provides potential for application in cancer malignancy determination, effects of cancer therapies, blood flow to organs, and assessments of therapies that influence blood flow to organs. Conventional ultrasound machines may be upgraded to operate in accordance with the invention by reprograrnming.

*****

It will be appreciated by those of ordinary skill in the art having the benefit of this disclosure that numerous variations from the foregoing illustration will be possible without departing from the inventive concept described therein. Accordingly, it is the claims set forth below, and not merely the foregoing illustration, which are intended to define the exclusive rights claimed in this application program.

Claims

5. CLAIMSWHAT IS CLAIMED IS:
1. A machine-executed method of generating a reference power value to represent the boundary between two different types of media in a region of a living organism, the method comprising: a) generating an image from a set of data generated by a scan of the region, the image comprising a plurality of pixels, each pixel having an associated power value; b) generating a cumulative power distribution curve representing the number of pixels with that power value or less in a coordinate system defined by an abscissa axis and an ordinate axis; c) fitting a line to the cumulative power distribution curve where the line intersects the cumulative power distribution curve at a first intersection point and a second intersection point; d) rotating the cumulative power distribution curve with the first and second points of intersection until the intersection points lie on the abscissa axis; e) fitting a parabola to a region surrounding the highest point on the rotated cumulative power distribution curve by the method of least squares, wherein the highest point represents 100 percent blood flow; and f) defining an abscissa value of the highest point of the parabola as the reference power value.
2. The method of claim 1 wherein the generating of an image comprises using an imaging machine selected from the group consisting of an ultrasound imaging machine, an x-ray imaging machine, and a nuclear magnetic resonance imaging machine.
3. The method of claim 2 wherein the cumulative power distribution curve function is defined by a first region from a medium of 100 percent blood and a second region from a tissue medium.
4. The method of claim 3 wherein the first intersection point is in the first region and the second intersection point is in the second region.
5. The method of claim 4 wherein the fitting of a line and the fitting of a parabola comprises using the method of least squares.
6. The method of claim 5 wherein the reference power value is used to set a green tag in an estimate of fractional moving blood volume.
7. The method of claim 5 wherein the reference power value is used to identify the boundary between the region of tissue and the region of blood.
8. A machine-executed method of generating a reference power value to represent the boundary between two different types of media in a region of a living organism, the method comprising: a) generating an image from a set of data generated by a scan of the region, the image comprising a plurality of pixels, each pixel having an associated power value; b) generating a cumulative power distribution curve representing the number of pixels with that power value or less in a coordinate, system defined by an abscissa axis and an ordinate axis; c) fitting a line to the cumulative power distribution curve where the line intersects the cumulative power distribution curve at a first intersection point and a second intersection point; d) constructing first and second tangent lines, where the first tangent line is tangent to the first intersection point and the second tangent line is tangent to the second intersection point; e) extending the first and second tangent lines until they define a boundary point; and f) defining an abscissa value of the boundary point as the reference power value.
9. The method of claim 8 wherein the generating of an image comprises using an imaging machine selected from the group consisting of an ultrasound imaging machine, an x-ray imaging machine, and a nuclear magnetic resonance imaging machine.
10. The method of claim 9 wherein the cumulative power distribution curve function is defined by a first region from a medium of 100 percent blood and a second region from a tissue medium.
1 1. The method of claim 10 wherein the first intersection point is in the first region and the second intersection point is in the second region.
12. The method of claim 1 1 wherein the fitting of a line comprises using the method of least squares.
13. The method of claim 12 wherein the reference power value is used to set a green tag in an estimate of fractional moving blood volume.
14. The method of claim 12 wherein the reference power value is used to identify the boundary between the region of tissue and the region of blood.
15. A machine-executed method of generating reference power values to represent the boundary between two different types of media in a region of a living organism, the method comprising: a) generating an image from a set of data generated by a scan of the region, the image comprising a plurality of pixels, each pixel having an associated power value; b) generating a first cumulative power distribution curve representing the number of pixels with that power value or less in a coordinate system defined by an abscissa axis and an ordinate axis; c) fitting a line to the first cumulative power distribution curve wherein the line intersects the first cumulative power distribution curve at a first intersection point and a second intersection point; d) manipulating the first cumulative power distribution curve to generate a reference power value.
16. The method of claim 15 wherein the manipulating of the first cumulative power distribution curve comprises: 1) rotating the first cumulative power distribution curve with the first and second points of intersection until the intersection points lie on the abscissa axis; 2) fitting a first parabola to a region surrounding the highest point on the rotated first cumulative power distribution curve by the method of least squares, wherein the highest point represents 100 percent blood flow; and 3) defining an abscissa value of the highest point of the first parabola as the first reference power value.
17. The method of claim 16 further comprising determining a second reference power value, the determining of a second reference power value comprising: 4) discarding power values that are less than the first reference power value;
5) generating a second cumulative power distribution curve;
6) fitting a second line to the second cumulative power distribution curve, wherein the second line intersects the second cumulative power distribution curve at a third intersection point and a fourth intersection point; 7) rotating the second cumulative power distribution curve with the third and fourth points of intersection until the intersection points lie on the abscissa axis; 8) fitting a second parabola to a region surrounding the highest point on the rotated second cumulative power distribution curve by the method of least squares, wherein the highest point represents 100 percent blood flow; and 9) defining an abscissa value of the highest point of the second parabola as the second reference power value.
18. The method of claim 16 further comprising determining a second reference power value, the determining of a second reference power value comprising:
4) discarding power values that are less than the first reference power value; and 5) finding an average for the remaining power values, wherein the average is defined to be a second reference power value.
19. The method of claim 15 wherein the manipulating of the first cumulative power distribution curve comprises:
1) constructing first and second tangent lines, where the first tangent line is tangent to the first intersection point and the second tangent line is tangent to the second intersection point;
2) extending the first and second tangent lines until they define a first boundary point;
3) defining a first reference power value as the abscissa value of the first boundary point;
0. The method of claim 19 further comprising determining a second reference power value, the determining of a second reference power value comprising:
4) discarding power values that are less than the first reference power value;
5) generating a second cumulative power distribution curve; 6) fitting a second line to the second cumulative power distribution curve, wherein the second line intersects the second cumulative power distribution curve at a third intersection point and a fourth intersection point;
7) constructing third and fourth tangent lines, wherein the third tangent line is tangent to the third intersection point and the fourth tangent line is tangent to the fourth intersection point;
8) extending the third and fourth tangent lines until they define a second boundary point; and
9) defining a second reference power value as the abscissa value of the second boundary point.
21. The method of claim 19 further comprising determining a second reference power value, the determining of a second reference power value comprising:
4) discarding power values that are less than the first reference power value; and
5) finding an average for the remaining power values, wherein the average is defined to be a second reference power value.
22. The method of claim 15 wherein the generating of an image comprises using an imaging machine selected from the group consisting of an ultrasound imaging machine, an x-ray imaging machine, and a nuclear magnetic resonance imaging machine.
23. The method of claim 22 wherein the cumulative power distribution curve is defined by a first region from a medium of 100 percent blood and a second region from a tissue medium.
24. The method of claim 23 wherein the first intersection point is in the first region and the second intersection point is in the second region.
25. The method of claim 24 wherein the fitting of a line comprises using the method of least squares.
26. The method of claim 25 wherein the reference power value is used to set a green tag in an estimate of fractional moving blood volume.
27. The method of claim 25 wherein the reference power value is used to identify the boundary between the region of tissue and the region of blood.
28. A machine-executed method of identifying an ultrasound power level representing one hundred percent blood flow in a region of a living organism, the method comprising: a) generating an image from a set of data generated by an ultrasound scan of the region, the image comprising a plurality of pixels, each pixel having an associated power value; b) generating a first cumulative power distribution curve representing the number of pixels with that power value or less in a coordinate system defined by an abscissa axis and an ordinate axis; c) fitting a first line to the first cumulative power distribution curve wherein the first line intersects the first cumulative power distribution curve at a first intersection point and a second intersection point; d) constructing first and second tangent lines, where the first tangent line is tangent to the first intersection point and the second tangent line is tangent to the second intersection point; e) extending the first and second tangent lines until they define a first boundary point; f) defining a first reference power value as the abscissa value of the first boundary point; g) discarding power values that are less than the first reference power value; h) generating a second cumulative power distribution curve; i) fitting a second line to the second cumulative power distribution curve, wherein the second line intersects the second cumulative power distribution curve at a third intersection point and a fourth intersection point; j) constructing third and fourth tangent lines, wherein the third tangent line is tangent to the third intersection point and the fourth tangent line is tangent to the fourth intersection point; k) extending the third and fourth tangent lines until they define a second boundary point; and 1) defining the abscissa value of the second boundary point as the ultrasound power level for one hundred percent blood flow.
29. An imaging system for measuring blood flow in a region of a living organism containing both blood vessels and other tissue, comprising: a) a scanning subsystem configured to generate a set of data, referred to as an image of the region, comprising a plurality of elements, referred to as pixels, each pixel having an associated power value; b) a display; c) a programmable processing system operatively coupled to the scanning subsystem and to the display; and d) a program storage device readable by the processing system and encoding a program of instructions including instructions for (1) performing operations as recited in a specified one of claims 1 through 28 to generate a reference power value to represent the boundary between the 100 percent blood and the tissue region, and (2) displaying the reference power value on the display.
30. A program storage device readable by the machine of a specified one of claims 1 through 28 and encoding instructions for performing the operations recited in the specified claim.
PCT/US1997/004871 1996-12-06 1997-03-26 Method and apparatus for cumulative distribution image boundary detection WO1998024363A1 (en)

Priority Applications (2)

Application Number Priority Date Filing Date Title
US76091796 true 1996-12-06 1996-12-06
US08/760,917 1996-12-06

Publications (1)

Publication Number Publication Date
WO1998024363A1 true true WO1998024363A1 (en) 1998-06-11

Family

ID=25060565

Family Applications (1)

Application Number Title Priority Date Filing Date
PCT/US1997/004871 WO1998024363A1 (en) 1996-12-06 1997-03-26 Method and apparatus for cumulative distribution image boundary detection

Country Status (1)

Country Link
WO (1) WO1998024363A1 (en)

Citations (6)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
US5148809A (en) * 1990-02-28 1992-09-22 Asgard Medical Systems, Inc. Method and apparatus for detecting blood vessels and displaying an enhanced video image from an ultrasound scan
US5322067A (en) * 1993-02-03 1994-06-21 Hewlett-Packard Company Method and apparatus for determining the volume of a body cavity in real time
US5339815A (en) * 1992-12-22 1994-08-23 Cornell Research Foundation, Inc. Methods and apparatus for analyzing an ultrasonic image of an animal or carcass
US5381791A (en) * 1992-03-10 1995-01-17 Siemens Medical Systems, Inc. Automatic indentification of anatomical features of interest from data acquired in nuclear medicine studies and automatic positioning of scintillation cameras to carry out such studies at optimal positions
US5489782A (en) * 1994-03-24 1996-02-06 Imaging Laboratory, Inc. Method and apparatus for quantum-limited data acquisition
US5494655A (en) * 1990-03-09 1996-02-27 The Regents Of The University Of California Methods for detecting blood perfusion variations by magnetic resonance imaging

Patent Citations (6)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
US5148809A (en) * 1990-02-28 1992-09-22 Asgard Medical Systems, Inc. Method and apparatus for detecting blood vessels and displaying an enhanced video image from an ultrasound scan
US5494655A (en) * 1990-03-09 1996-02-27 The Regents Of The University Of California Methods for detecting blood perfusion variations by magnetic resonance imaging
US5381791A (en) * 1992-03-10 1995-01-17 Siemens Medical Systems, Inc. Automatic indentification of anatomical features of interest from data acquired in nuclear medicine studies and automatic positioning of scintillation cameras to carry out such studies at optimal positions
US5339815A (en) * 1992-12-22 1994-08-23 Cornell Research Foundation, Inc. Methods and apparatus for analyzing an ultrasonic image of an animal or carcass
US5322067A (en) * 1993-02-03 1994-06-21 Hewlett-Packard Company Method and apparatus for determining the volume of a body cavity in real time
US5489782A (en) * 1994-03-24 1996-02-06 Imaging Laboratory, Inc. Method and apparatus for quantum-limited data acquisition

Similar Documents

Publication Publication Date Title
Ladak et al. Prostate boundary segmentation from 2D ultrasound images
Barry et al. Three-dimensional freehand ultrasound: image reconstruction and volume analysis
Brown et al. Registration of planar film radiographs with computed tomography
Guttman et al. Tag and contour detection in tagged MR images of the left ventricle
Tomazevic et al. 3-D/2-D registration of CT and MR to X-ray images
US5699799A (en) Automatic determination of the curved axis of a 3-D tube-shaped object in image volume
US4922915A (en) Automated image detail localization method
Lange et al. Augmenting intraoperative 3D ultrasound with preoperative models for navigation in liver surgery
US6052477A (en) Automatic technique for localizing externally attached fiducial markers in volume images of the head
US6389104B1 (en) Fluoroscopy based 3-D neural navigation based on 3-D angiography reconstruction data
US6370421B1 (en) Density modulated catheter for use in fluoroscopy based 3-D neural navigation
US6421454B1 (en) Optical correlator assisted detection of calcifications for breast biopsy
US5447154A (en) Method for determining the position of an organ
US7333648B2 (en) Feature quantification from multidimensional image data
US7231076B2 (en) ROI selection in image registration
US6368277B1 (en) Dynamic measurement of parameters within a sequence of images
US6542770B2 (en) Method of determining the position of a medical instrument
US7327865B2 (en) Fiducial-less tracking with non-rigid image registration
US6396940B1 (en) Optical correlator based automated pathologic region of interest selector for integrated 3D ultrasound and digital mammography
US7187792B2 (en) Apparatus and method for determining measure of similarity between images
Prager et al. Rapid calibration for 3-D freehand ultrasound
US7876939B2 (en) Medical imaging system for accurate measurement evaluation of changes in a target lesion
US6817982B2 (en) Method, apparatus, and product for accurately determining the intima-media thickness of a blood vessel
Krucker et al. Rapid elastic image registration for 3-D ultrasound
Xiao et al. Segmentation of ultrasound B-mode images with intensity inhomogeneity correction

Legal Events

Date Code Title Description
AL Designated countries for regional patents

Kind code of ref document: A1

Designated state(s): GH KE LS MW SD SZ UG AT BE CH DE DK ES FI FR GB GR IE IT LU MC NL PT SE BF

AK Designated states

Kind code of ref document: A1

Designated state(s): AL AM AT AU AZ BA BB BG BR BY CA CH CN CU CZ DE DK EE ES FI GB GE GH HU IL IS JP KE KG KP KR KZ LC LK LR LS LT LU LV MD MG MK MN MW MX NO NZ PL PT RO RU SD SE SG SI SK TJ TM TR TT UA UG US UZ VN YU AM AZ BY KG KZ MD RU TJ TM

CFP Corrected version of a pamphlet front page

Free format text: ADD INID NUMBER (63) "RELATED BY CONTINUATION (CON) OR CONTINUATION-IN-PART (CIP) TO EARLIER APPLICATION" WHICH WAS INADVERTENTLY OMITTED FROM THE FRONT PAGE

121 Ep: the epo has been informed by wipo that ep was designated in this application
NENP Non-entry into the national phase in:

Ref country code: CA

REG Reference to national code

Ref country code: DE

Ref legal event code: 8642

122 Ep: pct application non-entry in european phase