Image analysis – Applications – Biomedical applications
Reexamination Certificate
1999-10-29
2004-03-23
Boudreau, Leo (Department: 2621)
Image analysis
Applications
Biomedical applications
C378S054000
Reexamination Certificate
active
06711282
ABSTRACT:
BACKGROUND OF THE INVENTION
1. Field of the Invention
The present invention relates generally to image processing and specifically to a method for automatically generating a contour of a target bone based from a digital image.
2. Background of the Invention
Osteoporosis is a bone disease characterized by low bone mass and microarchitectural deterioration of bone tissue. This disease subjects a person to enhanced bone fragility and a consequent increase in fracture risk, particularly in the spine, hip and wrist. Osteoporosis is particularly common in postmenopausal women because their bone loss greatly exceeds that of men at this age. It has been estimated that at least 28 million Americans, 80% of whom are women, have a lower than normal bone mass and are at risk of having osteoporosis. In the United States alone, 10 million people already have osteoporosis and many women die each year from complications due to osteoporosis.
The assignee of the present invention has developed a proprietary bone density measurement system called OsteoGram Analysis. The OsteoGram Analysis system involves taking a standard X-ray of three fingers, along with a calibration wedge in the field of view by using existing and widely available standard X-ray equipment. The calibration wedge is utilized to adjust for differences among X-ray equipment, exposures, types of film, and the development process. The assignee of the present invention has a central specialized laboratory that receives these x-rays and provides a service to medical professionals of providing a bone mineral density (BMD) report based on the x-ray.
Although this service provides accurate reports, there has been a demand for an on-site solution that can be utilized by physicians or other medical professionals locally at their offices to generate the BMD report based on hand x-rays.
An important step is generating a contour of a target bone which is then used to calculate BMD. The lab service, described above, uses trained technicians to draw and identify the contour. Unfortunately, this approach to determining the contour of the bone requires and is dependent on the skill and training received by the technicians who administer this task. This approach involves cost, complexity, and training time which is not amendable to an on-site solution.
An alternative approach is to generate the contour of bone by employing some type of automatic contour extraction software program. An example of such an approach is described in U.S. Pat. No. 5,696,805, which is entitled “Apparatus and Method for Identifying Specific Bone Regions in Digital X-Ray Images.” Unfortunately, this approach suffers from several disadvantages.
First, the binarization approach to separate background information from desired objects, described in the '805 patent, is inadequate because there can be various inhomogeneities that exist within the actual hand image (e.g., radiographic shadow, scatter, inconsistent X-Ray field, etc.) that can render the binarization method inadequate. In many cases, a general approach is not robust enough to accurately separate the desired objects given the vast variability that exists among image acquisition and digitization.
Second, the skeletonizing technique, which is described in greater detail in U.S. Pat. No. 5,574,803, which is entitled “Character Thinning Using Emergent Behavior of Populations of Competitive Locally Independent Processes,” is complex and difficult to implement, thereby adding to the development costs and time in implementing this approach.
Third, this approach uses a grayscale analysis to determine the location of joint spaces, which can be inaccurate in cases where there are fused joints or overlapping joint areas.
Fourth, this approach does not address or describe how the bottom and top bone edges are extracted, particularly in the joint space areas, which is an important step in accurately generating the contour of the bone.
Fifth, this approach does not provide an architecture or systematic approach, thereby making this approach difficult to incorporate, customize, develop, refine, or test new and different methods.
In addition, this approach does not allow for user verification of the segmentation. For example, if faulty or inaccurate segmentation in the form of errors should occur for visible reasons (e.g., severe arthritis, fused joints, bone spurs, bone calcifications, image artifacts, etc.) or invisible reasons (e.g., radiographic shadow, x-ray scatter and noise, poor contrast, etc.), then these errors can propagate through the entire system described in the '805 patent. These errors render the final output useless. Furthermore, there is no mechanism for a user to provide feedback to the program in order to improve intermediate results.
Accordingly, there remains a need for a method for automatically generating a contour of a target bone based on a digital image that departs significantly from existing methods and that overcomes the disadvantages set forth previously.
SUMMARY OF THE INVENTION
Accordingly, it is an object of the present invention to automatically generate a contour of a target bone.
It is a further object of the present invention to automatically generate a contour of a target bone and receive user input to refine the contour of the target bone.
It is yet another object of the present invention to provide an automatic segmentation module that utilizes shape analysis to determine the location of joint edge points.
It is yet another object of the present invention to provide an automatic segmentation module that utilizes candidate selection to determine the location of joint edge points.
It is yet another object of the present invention to provide an automatic segmentation module that utilizes minimum flow analysis to determine the top and bottom edges of the target bone.
It is yet another object of the present invention to provide an automatic segmentation module that has an object oriented architecture that allows for feedback verification, user intervention, and flexibility to incorporate different methods.
In accordance with the present invention, the foregoing objects are met in an automatic segmentation module with an open architecture that automatically generates a contour of a target bone based on a digital image. In one embodiment, the digital image can include an extremity, such as a hand with three digits and a calibration wedge. The automatic segmentation module receives the digital image as an input and automatically segments the target bone by utilizing an object-oriented approach. This approach extracts intermediate target objects and uses these intermediate target objects to refine or limit the search space in order to extract the final target object.
REFERENCES:
patent: 4721112 (1988-01-01), Hirano et al.
patent: 5291537 (1994-03-01), Mazess
patent: 5335260 (1994-08-01), Arnold
patent: 5365564 (1994-11-01), Yashida et al.
patent: 5426709 (1995-06-01), Yoshida et al.
patent: 5533084 (1996-07-01), Mazess
patent: 5574803 (1996-11-01), Gaborski et al.
patent: 5673298 (1997-09-01), Mazess
patent: 5696805 (1997-12-01), Gaborski et al.
patent: 5712892 (1998-01-01), Weil et al.
patent: 5852647 (1998-12-01), Schick et al.
patent: 5898753 (1999-04-01), Schick et al.
patent: 5910972 (1999-06-01), Ohkubo et al.
Brent J. Lui, Ricky K. taira, HyeonJoon Shim, Patricia keaton, “Automatic Segmentation of bones from digital hand radiographs,” SPIE vol. 2434 0-8194-1782-3/95, pp. 659-669.*
Hyeonjoon Shim, Brent J. Liu, Ricky K. Taira, Theodore Hall, “Object-oriented approach towards the automatic segmentation of bones from pediatric hand radiographs,” SPIE vol. 3034 0277-786X/97, pp. 95-105.
Brent J. Liu, Ricky K. Taira, HyeonJoon Shim, Patricia Keaton, “Automatic Segmentation of bones from digital hand radiographs,” SPIE vol. 2434 0-8194-1782-3/95, pp. 659-669.
Bi Xiaoli
Duff Eric
Edelstein David
Liu Brent J.
Shim Hyeonjoon
Akhavannik Hussein
Boudreau Leo
Compumed, Inc.
Ho Eric
LandOfFree
Method for automatically segmenting a target bone from a... does not yet have a rating. At this time, there are no reviews or comments for this patent.
If you have personal experience with Method for automatically segmenting a target bone from a..., we encourage you to share that experience with our LandOfFree.com community. Your opinion is very important and Method for automatically segmenting a target bone from a... will most certainly appreciate the feedback.
Profile ID: LFUS-PAI-O-3227297