Image analysis – Applications – Biomedical applications
Utility Patent
1998-11-04
2001-01-02
Mehta, Bhavesh (Department: 2721)
Image analysis
Applications
Biomedical applications
C128S126100, C345S419000, C345S424000
Utility Patent
active
06169817
ABSTRACT:
BACKGROUND OF THE INVENTION
The present invention is directed to a system of and method for reconstruction and visualization of an object from raw image data of the object in four dimensions (three dimensions of space and one of time), using already known characteristics of the object to analyze the raw image data. More particularly, the present invention is directed to a system of and method for visualizing an object from a time sequence of bitmapped images of the object. In one embodiment, the present invention is more specifically directed to such a system and method for use in imaging MRI (magnetic resonance imaging) data of human body parts. The distinct organs, tissues and bones are reconstructed in four dimensions in virtual (digital) space, creating a “digital clone” or virtual reality replica of the moving human body.
The ability to evaluate the motion patterns of human joints, including both the bones and the soft tissue structures, would greatly increase understanding of the development and staging of such conditions as osteoarthrosis. Although there is a large body of information about the biomechanics and kinematics of the anatomical structures of human joints, the information is based mostly on cadaver studies and on two-dimensional image-based studies. So far, there have been no successful systems that can display in vivo three-dimensional joint kinematics.
While it is known in the art to reconstruct and visualize high-contrast anatomical structures in three spatial dimensions, the resulting reconstruction is static and provides no time-based information. Many problems to be diagnosed and treated are related to abnormal motion kinematic events in complex structures such as joints; therefore, imaging in four dimensions (both time and space) is desirable.
The best current source of raw image data for observation of a complex soft tissue and bone structure is magnetic resonance imaging (MRI). However, MRI often introduces the following technical challenges. Many of the anatomical structures to be visualized require high resolution and present low contrast, since many of the musculoskeletal structures to be imaged are small and intricate. MRI involves the use of local field coils to generate an electromagnetic field; such local field coils form a non-uniform illumination field. MRI images can also be noisy. Any technique for reconstruction and visualization that uses MRI image data should meet those challenges.
Although there has been some research in the area of deformable motion tracking and analysis, most of that research has concentrated on the time evolution of a single structure in the image based on simple parametric deformable models. The kinematic analysis of a joint involves the motion of many structures, thus making the conventional approaches unsuitable.
Moreover, the known techniques for 3D reconstruction and visualization offer the following disadvantages. First, such known techniques are too computationally intensive to be used practically for a time sequence of images. Second, most known algorithms for 3D reconstruction and visualization are supervised (i.e., require operator intervention) and rely on the expert knowledge of a radiologist or another such person. That supervision limits the range of applications of those algorithms to some specific, anatomically simple structures such as bones, the heart, and the hippocampus. On the other hand, the successfully unsupervised techniques in the prior art are specific to one organ and are thus limited in the range of anatomical structures that they can handle.
Examples of supervised techniques are taught in the following references:
W. E. Higgins et al, “Extraction of left-ventricular chamber from 3D CT images of the heart,”
Transactions on Medical Imaging
, Vol. 9, no. 4, pp. 384-394, 1990;
E. A. Ashton et al, “A novel volumetric feature extraction technique with applications to MRI images,”
IEEE Transactions on Medical Imaging
, Vol. 16, pp. 365-371, 1997; and
M. W. Hansen et al, “Relaxation methods for supervised image segmentation,”
IEEE Trans. Patt. Anal. Mach. Intel.
, Vol. 19, pp. 949-962, 1997.
Examples of unsupervised techniques are taught in the following references:
W. M. Wells III et al, “Adaptive segmentation of MR data,”
IEEE Transactions on Medical Imaging
, pp. 429-440, 1996; and
J. Rajapalse et al, “Statistical approach to segmentation of single-channel cerebral MR images,”
IEEE Transactions on Medical Imaging
, Vol. 16, no. 2, pp. 176-186, 1997.
The inventors have presented related concepts in Tamez-Pe{tilde over (n)}a et al, “Automatic Statistical Segmentation of Medical Volumetric Images,”
IEEE Computer Vision and Pattern Recognition
98. All of the references cited above are hereby incorporated by reference in their entirety into the present disclosure.
SUMMARY AND OBJECTS OF THE INVENTION
An object of the present invention is to provide a system of and method for 4D (space and time) kinematic reconstruction and visualization of body tissue or the like.
Another object of the present invention is to provide a system of and method for reconstruction and visualization that incorporates motion tracking.
Still another object of the present invention is to provide a system of and method for reconstruction and visualization that allow an unlimited number of tissue segments at the same time.
Yet another object of the present invention is to provide a system of and method for reconstruction and visualization that take into account the biomechanical properties of the tissue being reconstructed and visualized.
Still another object of the present invention is to provide a system of and method for visualization and reconstruction that incorporate a database of such biomechanical properties.
Another object of the present invention is to provide a system of and method for visualization and reconstruction that produce a finite element biomechanical model.
An additional object of the present invention is to provide a system of and method for visualization and reconstruction that produce a “digital clone” or virtual reality replica of the human body that can be acted on in virtual space and whereby the body parts respond in a realistic manner in virtual space to applied forces.
To achieve these and other objects, the present invention is directed to a system and method that implement the following technique. One image in the sequence of images, typically the first, is segmented through statistical techniques to identify the tissues and structures shown in the image. Then, instead of doing the same thing for all of the images, the segmenting is carried over to all of the images by estimating the motion of each of the tissues and structures. The estimation takes into account known properties of the tissues and structures expressing the known properties in terms of spring elastic constants.
The segmenting can be done in the following manner. First, a time sequence of 3D images is formed, as by MRI, to obtain raw image data. The first image is divided or segmented into regions, each of the regions corresponding to a tissue or structure of the body part being imaged. The statistical technique for segmenting involves estimating a local mean value and a local variance for the image in each voxel (three-dimensional pixel). The voxels are joined into regions if their estimated local mean values and estimated local variances are sufficiently similar. Thus, regions are defined corresponding to the various tissues or structures, so that tendons, bones, and the like can be identified.
Once the segmenting is done on the first image, it is carried over to the other images to save the computational power that would be used to segment all of the images in the same manner. That carrying over uses known motion characteristics of the various components and expresses them as spring constants. For example, the characteristics of the connection between a tendon and a bone can be expressed by imagining a spring that connects the two and giving its spring constant. Those known motion characteristics can be used to estim
Parker Kevin J.
Tamez Pe{tilde over (n)}a Jose
Totterman Saara Marjatta Sofia
Blank Rome Comisky & McCauley LLP
Mehta Bhavesh
Patel Kanji
University of Rochester
LandOfFree
System and method for 4D reconstruction and visualization does not yet have a rating. At this time, there are no reviews or comments for this patent.
If you have personal experience with System and method for 4D reconstruction and visualization, we encourage you to share that experience with our LandOfFree.com community. Your opinion is very important and System and method for 4D reconstruction and visualization will most certainly appreciate the feedback.
Profile ID: LFUS-PAI-O-2479773