This invention relates to medical diagnostic ultrasound systems and, in particular, to ultrasound systems which perform three dimensional (3D) color flow imaging.
It is common practice in ultrasonic diagnostic imaging to visualize the interior surface of structures of a human body, such as the ventricles of the fetal heart, using 3D volume rendered imaging. Using a 3D data acquisition technique known as Spatial Temporal Image Correlation (STIC), the dynamics of the fetal heart can be captured as a series of volumes, each captured as a Cineloop® of consecutive image frames. It is also common practice to capture 3D volumes of color flow data and form a 2D image projection of that data volume using various 3D volume rendering techniques. The 3D color flow data can then be viewed as a cross-sectional plane, or as a 3D image using various 3D rendering methods. Furthermore, the 3D color flow data can be acquired using the STIC technique to capture the hemodynamics of the fetal heart.
Another known technique for fetal imaging is known as “invert imaging.” In invert imaging, the conventional grayscale range which generally shows structures in the body which return strong echoes as brightly displayed and anechoic structures such as blood which return little echo energy as dark, in an inverted grayscale range. This reversal of the grayscale range results in the blood inside of vessels as shown brightly lighted with the tissue of the surrounding vessels dimly displayed or invisible, thereby highlighting blood pools and vessel blood flow. See, for instance, U.S. Pat. No. 6,117,080 (Schwartz), which applies this technique in the detection of fluid-filled cysts.
The use of the fetal STIC technique in combination with invert imaging has provided clinicians with new insight into the structure of the fetal heart. This is important because the clinician is often presented with the problem of assessing the fetal heart for normal or abnormal formation and function. The conventional 3D invert image, by itself, does not provide any information about the hemodynamics of the heart. For that, the clinician must turn instead to another image, typically a cross-sectional color flow slice through the volume to assess blood flow velocity, and mentally correlate the flow of the cross-sectional color flow image with the 3D grayscale image formed with the invert technique. Accordingly it would be desirable to provide the clinician with a single imaging technique which simultaneously provides the vascular flow path information of an inverted image and the flow velocity information of the color flow slice.
In accordance with the principles of the present invention, 3D grayscale data is combined with 3D color flow data to allow them to be visualized together. In an illustrated example of the invention, the 3D projection of the surface rendering of the color volume data is combined with the 3D projection of the surface rendering of the “inverted” grayscale volume data to produce an image of the two together. In this example the process of combining the two 3D projections compares the value of the grayscale projection data at a given pixel location with the value of the color projection data at the same pixel location. If the grayscale value is below a certain threshold, only the grayscale value is used for the image pixel in the combined image. If the grayscale value is above that threshold, the grayscale value is added to the color value and that new value is used for the image pixel in the combined image. Other, image data algorithms can also be used to enhance features of the image.
In the drawings:
Referring first to
The partially beamformed signals produced by the microbeamformer 12a are coupled to a main beamformer 20 where partially beamformed signals from the individual patches of elements are combined into a fully beamformed signal. For example, the main beamformer 20 may have 128 channels, each of which receives a partially beamformed signal from a patch of 12 transducer elements. In this way the signals received by over 1500 transducer elements of a two dimensional array can contribute efficiently to a single beamformed signal. The beamformed signals are coupled to a signal processor 22 where they may undergo additional enhancement such as speckle removal, signal compounding, harmonic separation, filtering, multiline interpolation and processing, and noise elimination.
The processed signals are coupled to a B mode processor 26 and a Doppler processor 28. The B mode processor 26 employs amplitude detection for the imaging of structures in the body such as muscle, tissue, and blood cells. B mode images of structure of the body may be formed in either the harmonic mode or the fundamental mode. Tissues in the body and microbubbles both return both types of signals and the harmonic returns of microbubbles enable microbubbles to be clearly segmented in an image in most applications. The Doppler processor processes temporally distinct signals from tissue and blood flow for the detection of motion of substances in the image field including blood cells, tissue, and microbubbles. The Doppler processor operates on ensembles of temporally distinct samples from each location in the volume being imaged to produce an estimate of Doppler power, velocity, acceleration, or variance at each location in the volumetric region, the same volumetric region which is the source of the B mode signals. Different transmit signals may be used for B mode and Doppler returns or the same signals may be used by both processors as described in U.S. Pat. No. 6,139,501 (Roundhill et al.) The Doppler processor 28 can operate on the I,Q quadrature data typically produced by the signal processor 22 and the B mode processor 26 can operate on the same data in the form of (I2+Q2)1/2. The structural B mode signals are coupled to a grayscale mapping processor 32 which converts the B mode signals to a range of grayscale values. The flow signals from the Doppler processor 28 are coupled to a color mapping processor 34 which similarly converts the flow signals to a range of color values. When the flow signals are velocity-related signals for color flow imaging the range of color values corresponds to a range of flow velocities, for instance. Other Doppler modes such as power Doppler, acceleration, and variance may be used if desired. The mapping processors may implement grayscale and color ranges selected by the user and may be constructed as lookup tables. When the ultrasound system of
In accordance with the present invention the grayscale 3D data set and the 3D flow data set are each volume rendered to form a 2D display of each 3D volume of data by volume renderers 42 and 44, respectively. In practice one volume renderer may be used which is multiplexed to render one data set and then the other. Volume rendering is well known and is described in U.S. Pat. No. 5,720,291 (Schwartz), for instance. See also U.S. Pat. No. 6,530,885 (Entrekin et al.) Volume rendering can produce projections of a 3D volume from a series of different look directions and the user can then sequence through the look directions to view the 3D volume from different perspectives, a display format known as kinetic parallax. The volume renderers 42 and 44 can operate on image data in either rectilinear or polar coordinates as described in U.S. Pat. No. 6,723,050 (Dow et al.) In further accord with the present invention the volume rendered, “inverted” grayscale data produced by the volume renderer 42 and the volume rendered flow data produced by the volume renderer 44 are blended together by a pixel comparator and 3D image merge processor 50, which allows the flow path of blood vessels and heart chambers to be visualized together with the fluid motion characteristics of the flow path or chamber. One way to do this is to compare the B mode and flow data at each point in the image with each other or a threshold. For instance, one technique is to compare the grayscale value at an image location with a threshold and if the grayscale value is below the threshold, the grayscale value alone is used for display. But if the grayscale value exceeds the threshold, the grayscale and flow values are summed, resulting in a display pixel at that location which has been tinted with the flow value such as flow velocity. Alternatively, the flow value is first compared with a threshold and used for display if it exceeds the threshold. If the flow value does not exceed the threshold the grayscale value is used for display at that location. In either case, the thresholds may be controlled and set by the user from the control panel 38, if desired. The result of this processing is that a single volume rendered (3D) image is used which contains characteristics of both of the initial volume renderings. The resultant merged image is coupled to a Cineloop buffer 36 where it may be displayed with a sequence of similarly processed imaged to visualize the hemodymanics of the heart, for instance. The images to be displayed are coupled to an image processor 30 for display on an image display 40.
To recap this processing, the formation of the combined grayscale and color 3D image projections is realized in a series of processing steps. These steps involve first the formation of the individual grayscale and color data projections, and then the final step of combining or compositing them into a single image for display. The formation of the grayscale image projection is realized through the rendering of grayscale data using an inverted grayscale map from the conventional map, where the first step in the process is to reverse, or invert the intensity of the individual voxels of the volume of data to be visualized. This inversion is such that it causes bright voxels to become dark, and dark voxels to become bright. Subsequent to this inversion step, a conventional 3D rendering technique such as the ray-cast method previously described is used to create a 2D image which is a projection of the 3D data as observed from a given viewpoint. Preferably, this rendering method is configured to find and show “surfaces” in the data, surfaces meaning generally the transition from low intensity (dark) voxels to higher intensity (bright) voxels as seen by a “ray” traveling from the observer into the volume of data. Because of the first step of inverting the intensities of the voxel data, the surfaces found and displayed during the rendering process are the equivalent to the interior surfaces of tissue which oppose anechoic, or normally dark regions present in the volume data. If the volume data includes a fetal heart, for example, the surfaces found and displayed would correspond to the interior surfaces of the ventricles of the heart and associated blood vessels connecting to it. The resulting 2D image is composed of MXN pixels, with each pixel consisting of a red (R), green (G), and blue (B) value. Typically these RGB values will all be equal to one another so that the displayed pixel will have a neutral (grey) color, although other tints can be used.
The formation of a color flow image projection is realized through conventional 3D rendering methods, such as the ray-cast method, to find and display the surface of targets that were previously detected by the ultrasound system to have a Doppler shift associated with them (moving blood cells) within the volume data, as viewed from a particular viewpoint. The resulting 2D image is composed of MXN pixels, each pixel consisting of a red (R), blue (B), and green (G) value. The RGB values of the pixels will be such that the resulting color displayed at each pixel location will correspond to the direction and velocity of the blood cells at that location as determined by a color map that relates velocity and direction as determined by the Doppler shift of the received echoes to displayed color.
The combined color and grayscale 2D projections of the 3D data created above are combined, or composited, in a non-linear fashion where each pixel of the combined image is formed as a combination of the pixel (voxel) data from the corresponding location in the grayscale and color images. There are several ways to combine the pixel data from the two images. One way is to first compare the grayscale pixel value to a selected threshold. If the pixel value falls below an adjustable threshold, only that RGB pixel value from the grayscale image is used for the combined image pixel, that is, only the grayscale data is displayed. On the other hand, if the pixel value from the grayscale data exceeds the adjustable threshold, then that pixel value is summed with the pixel value of the color image (R, G, and B summed individually) at that location, resulting in a grayscale pixel value that has been tinted with the corresponding velocity found at that same location. If desired, a range check can be performed to clamp the summed pixel value to full-brightness. This compositing method is opposite to the conventional process where color data is only displayed in the absence of grayscale data.
Another way to composite the grayscale and color images is to use an additional adjustable threshold for the values of the color pixel data such that the velocity of the color data is also taken into consideration. In this case, it may be desired to show the color pixel data regardless of the grayscale data if the velocity of the color data exceeds the adjustable threshold.
Yet another factor that can influence the compositing process is to consider the location in the volume from which the grayscale surface and the color surface portrayed by the grayscale and color image pixels originated. When using the ray-cast method of volume rendering, the final result is a two dimensional projection in which the depth dimension is normally lost. However, the depth location in the original volume data set from which the pixel originated can be determined by keeping track of the distance along the ray where the surface was found for each of the grayscale and color surfaces. Then, in addition to comparing the grayscale and color pixel values as part of the compositing process, the depths along the ray where each pixel was encountered can also be considered. For instance, when the grayscale pixel value in the grayscale rendering originated from a shallower depth than the corresponding color pixel value in the color flow rendering, the two values should not be merged as they relate to different locations in the volumetric region. In such case, if the grayscale value exceeds the specified brightness threshold, but the depth along the ray was different between the color pixel and grayscale pixel by some adjustable threshold, only the grayscale pixel value alone will be used for the combined image pixel. This additional factor prevents the merging of rendered data which does not spatially correlate and thus should not be combined.
A simplified flowchart of a process of the present invention is shown in
Filing Document | Filing Date | Country | Kind | 371c Date |
---|---|---|---|---|
PCT/IB2007/054044 | 10/4/2007 | WO | 00 | 4/7/2009 |
Number | Date | Country | |
---|---|---|---|
60829353 | Oct 2006 | US |