The present invention relates to the diagnostic imaging arts. It finds particular application in conjunction with a CT imaging system and will be described with particular reference thereto. However, it is to be appreciated that the present invention is applicable to a wide range of diagnostic imaging modalities.
In the field of medical imaging, it is often necessary to recognize changes of the object to be examined on the basis of different images of the same object. It is often desired that the images acquired at different instances should enable the physician to recognize which of the changes appearing in the imaged object are due to natural motions and deformations and which changes can be attributed to pathological changes such as, for example, tumor growth. Images of a subject which have been formed before and after an operation or treatment are routinely compared so as to assess the result of the treatment.
Typically, the images, which are formed at different instances by the same or different modalities, have to be registered by the means of scaling, rotating and the like to have the position and shape of the organs coincide. Rigid transformations are defined as geometrical transformations that preserve distances. The rigid transformations also preserve straightness of lines and all non-zero angles between straight lines. The rigid transformations are typically composed of translations and rotations. When the bending of joints and the respiratory motion constitute flexible or non-rigid motions, the anatomical object to be examined cannot be shifted to its original position by rigid transformations such as rotation and translation. In this case, an elastic registration is typically used.
In elastic registration, the image is modeled as an elastic body and the similarity between points or features in the two images act as external forces, which stretch the body. Elastic registration of images is used for a wide variety of clinical applications where images that have been acquired at different times, with different modalities, or for different patients need to be aligned with one another. The examples of images requiring elastic transformation include tumor diagnosis, surgery and treatment, where the images are typically taken at different modalities to show different aspects of the tumor, taken at different times to compare effects of pre-intervention and post-intervention images, or being matched with the anatomical atlases derived from cohorts studies.
Typically, the images are first segmented to designate a region of interest to guide the registration. After the images are segmented and common points to the two images are established, the images are registered by using the rigid transformation, as step one; and the elastic transformation, as step two. Because of the complexity of the elastic transformations, sometimes the image structures are not properly aligned. In this case, it is desirable that the user manually corrects the registration, following the elastic transform registration, by introducing deformations to the segmented surfaces in the 3D images. Generally, the manual correction of the 3D images registration is difficult as the 3D datasets include large volumes of data to be transformed following the manual deformation.
The present application contemplates a new and improved method and apparatus which overcomes the above-referenced problems and others.
In accordance with one aspect of the present invention an apparatus for diagnostic imaging is disclosed. A first memory supplies a first diagnostic image. A second memory supplies a second diagnostic image. A registration routine automatically registers the first and second diagnostic images from the first and second image memories. A display concurrently displays at least a corresponding pair of 2D slices of the first and second registered diagnostic images. A means manually transforms at least one of the currently displayed 2D slices corresponding to one of the first and second registered diagnostic images.
In accordance with another aspect of the present invention a method of diagnostic imaging is disclosed. A first diagnostic image of a selected region is supplied. A second diagnostic image of the selected region is supplied. The first and second diagnostic images are automatically registered. A corresponding pair of 2D slices of the first and second registered diagnostic images is concurrently displayed. At least one of the currently displayed 2D slices corresponding to one of the first and second registered diagnostic images is manually transformed.
One advantage of the present invention resides in computational efficiency.
Another advantage resides in efficiency of correction of misregistered images.
Another advantage resides in real time display of corrected images.
Still further advantages and benefits of the present invention will become apparent to those of ordinary skill in the art upon reading and understanding the following detailed description of the preferred embodiments.
The invention may take form in various components and arrangements of components, and in various steps and arrangements of steps. The drawings are only for purposes of illustrating the preferred embodiments and are not to be construed as limiting the invention.
With reference to
Image data from the hospital archive or from another storage medium 16 of the same region of the same subject is retrieved and stored in an archived 3D volumetric image memory 18 (image B). Of course, both the current and archive 3D image memories 14, 18 may be parts of a common storage medium.
With continuing reference to
An aligning means or process 38 registers the images A, B for a concurrent display on one or more monitors or displays 40. More specifically, an affine transform means 42 performs a first step of the aligning process 38 and approximately aligns images A, B by determining a misalignment between point landmarks in the current and archived 3D images A, B. Specifically, the affine transform means 42 searches for the most distinguished anatomical features in the segmented areas of the images A, B such as characteristic portions of the body around the region of interest, e.g. unique locations on the skull or the vertebrae, and determines an affine transform between the misaligned landmarks. Alternatively, the affine transform means 42 searches for the fiducials or imagable markers that have been affixed to the subject closely adjacent the region of interest. When such common points are determined, the affine transform means 42 applies appropriate algorithms, known in the art, to align the images A, B. In one embodiment, the affine transform means 42 determines nine rotational components about three orthogonal axes and three translational components along the three axes that define the registration error. Optionally, a scaling parameter can also be determined.
An elastic transform means 44 performs a second step of the aligning process 38 by a use of a point based elastic registration. More specifically, the elastic transform means 44 determines misalignment between the landmarks caused by non-rigid motions and the like and applies a closed-form elastic transformation to the misaligned landmarks. More specifically, the closed-form Gaussian elastic transformation uses the Gaussian-shaped forces centered at the positions of the landmarks to elastically deform the images A, B in a way that the prescribed landmark correspondences (displacements) are preserved.
The elastic transform means 44 preferably applies an elastic transform operator:
The images A, B, aligned by the affine transform means 42 and the elastic transform means 44, are stored in an aligned images memory 46.
A video processor 50 formats the aligned images A, B for display on the monitor 40 of a workstation 52 such that corresponding first and second sets of 2D slices of the aligned images A, B are displayed concurrently. A user manipulates the displayed slices using the workstation 52 which includes a CPU processor or hardware means 54 and a software means 56 for carrying out the necessary image processing functions and operations. The workstation 52 preferably includes one or more input devices 58 by the use of which the user can selectively control the workstation 52 and/or the scanner 10.
With continuing reference to
With continuing reference to
In one embodiment, the Gaussian pull tool 74 pulls a Gaussian shaped distortion (or other functional shape the smoothly transitions from 1 to 0) but derives the distance that the distortion is pulled from the distance of the mouse position from the 2D image plane. The 2D surface is pulled directly to the mouse position enabling smooth drawing, rather than having to click up and down on the mouse to grab and stretch the organ. Although the Gaussian is applied to the image in the displayed slices, it can affect the other two orthogonal slices if it is applied near the crossing point. Moreover, the Gaussian deformation also affects neighboring parallel slices. However, the neighboring slices are not modified at the present time. Rather, the deformation parameters in other planes are saved and the modification to the surface in each neighboring plane is made when and if such neighboring plane is called up for display.
With continuing reference to
The reformatted slices are stored in a data memory 90. Preferably, the reformatted slices are stored in a cache-conscious way to accelerate the reinspection if so requested.
The user, through the keyboard or other input device 58 controls a stepping means 92 which causes the video processor 50 to withdraw and display corresponding 2D slices from the data memory 90 on the monitor 40. When the user changes the view by scrolling through the slices, the corresponding manually transformed regions are preferably calculated and updated on the fly. In one embodiment, an updating means 94 pre-computes slices adjacent to the currently displayed corrected slices without waiting for the user interaction. If the user elects, the reformatted slices to become part of the permanent record for storage in the electronic archives, the 3D image is automatically updated. Preferably, the update of the 3D image in accordance with manual transformations of the 2D slices is done at the session closing or at the dead time. For example, the user activates a “save” option (not shown) on the monitor 40 which action initiates saving and updating of the 3D image.
In one embodiment, the corresponding slices of the images A, B are superimposed. The user uses the manual tools 72 to deform surfaces in one or both images A, B to align the image A, B with one another.
Although described with particular reference to CT scanner imaging, it is to be appreciated that this technique is also applicable to magnetic resonance images, PET images, SPECT images, and other three-dimensional diagnostic images. Moreover, the images being registered may be from mixed modalities. For example, a CT image can be registered using this technique with a PET image. When mixing modalities, care should be taken to assure that the features are defined in both imaging modalities or appropriate adjustment made. It is to be appreciated that this technique is applicable to studies of a variety of organs such as the colon, the liver, and other non-rigid organs. Moreover, this technique is also applicable to rigid portions of the body such as the head.
The invention has been described with reference to the preferred embodiments. Modifications and alterations will occur to others upon a reading and understanding of the preceding detailed description. It is intended that the invention be construed as including all such modifications and alterations insofar as they come within the scope of the appended claims or the equivalents thereof.
This application claims priority under 35 U.S.C. §371 from International PCT application Ser. No. PCT/IB05/053628 filed Nov. 4, 2005 which claims priority to U.S. provisional application Ser. No. 60/628,685 filed Nov. 17, 2004, which is incorporated herein by reference.
Filing Document | Filing Date | Country | Kind | 371c Date |
---|---|---|---|---|
PCT/IB2005/053628 | 11/4/2005 | WO | 00 | 5/16/2007 |
Publishing Document | Publishing Date | Country | Kind |
---|---|---|---|
WO2006/054191 | 5/26/2006 | WO | A |
Number | Name | Date | Kind |
---|---|---|---|
5633951 | Moshfeghi | May 1997 | A |
5682886 | Delp et al. | Nov 1997 | A |
5970182 | Goris | Oct 1999 | A |
6539127 | Roche et al. | Mar 2003 | B1 |
6728424 | Zhu et al. | Apr 2004 | B1 |
6754374 | Miller et al. | Jun 2004 | B1 |
7106891 | Wyman et al. | Sep 2006 | B2 |
7627158 | Hay | Dec 2009 | B2 |
20020097901 | Xu et al. | Jul 2002 | A1 |
20030063788 | Boland et al. | Apr 2003 | A1 |
20030128890 | Roesch et al. | Jul 2003 | A1 |
20030181808 | McKinnon | Sep 2003 | A1 |
20030233039 | Shao et al. | Dec 2003 | A1 |
20050013471 | Snoeren et al. | Jan 2005 | A1 |
20060204064 | Hay | Sep 2006 | A1 |
Number | Date | Country | |
---|---|---|---|
20080273779 A1 | Nov 2008 | US |
Number | Date | Country | |
---|---|---|---|
60628685 | Nov 2004 | US |