Image analysis – Applications – Biomedical applications
Reexamination Certificate
1999-06-07
2002-09-17
Johns, Andrew W. (Department: 2721)
Image analysis
Applications
Biomedical applications
Reexamination Certificate
active
06453058
ABSTRACT:
BACKGROUND
1. Technical Field
The present invention relates generally to computer-assisted diagnosis (CAD) and, in particular, to a CAD method using correspondence checking and change detection of salient features in digital images.
2. Background Description
Computer-assisted diagnosis is an important technology in many different clinical applications. However, one of the more prevalent clinical applications for computer-assisted diagnosis is in the detection of breast cancer in women. According to the American Cancer Society, breast cancer is the most common cancer among women, other than skin cancer. It is the leading cause of death among women aged 40 to 55. There are approximately 179,000 new cases of breast cancer in the United States each year and about 43,500 deaths from the disease.
While there are presently no means for preventing breast cancer, early detection of the disease prolongs life expectancy and decreases the likelihood of the need for a total mastectomy. Accordingly, the American Cancer Society recommends that all women aged 40 and older should have a mammogram every year.
Human reading of x-ray mammograms is seldom done in isolation. For example, diagnostic findings from a mammogram are confirmed often after comparing them with those on another mammogram. Change detection may be attempted between images of the same breast in the same view taken at different times (temporal), or between images of the left and right breasts taken at about the same time (bilateral).
As digital storage of mammograms becomes more widespread, it will become increasingly common to do change detection by automatic means in routine screenings. There is already some work in Computer Aided Diagnosis (CADx) for mammography that attempts to perform change detection, e.g., for the identification of masses. Typically with these methods the images first need to be registered. Registering images refers to the process of identifying the correspondences between pixels of the images. Registration methods used in the context of mammography include lining up breast outlines, estimating the correspondence matrices using a few manually specified correspondence points, and so on. Registration is difficult for mammograms because of the differential displacement of soft tissue and structures between images, local and global changes in the appearance of the breast, subtlety of diagnostic findings, and varying image acquisition parameters such as compression and exposure.
Moreover, humans do not seem to accomplish change detection through pixel by pixel comparisons between images. Rather, they are able to quickly detect salient features in the two images and then compare these features based on their location, size and other visual characteristics, while generally ignoring the remainder of the image.
Accordingly, it would be desirable and highly advantageous to have a computer assisted diagnosis method that avoids the problems associated with pixel-based registration methods.
SUMMARY OF THE INVENTION
The present invention is directed to a CAD method using correspondence checking and change detection of salient features in digital images. Changes are detected directly from feature characteristics without resorting to exact registration or exact matching.
In one aspect of the present invention, a computer assisted diagnosis method for automatically detecting changes of salient features in a first and a second digital image includes the step of converting the first and the second digital image into a first and a second relational attribute graph, respectively. Each graph comprises nodes and arcs. Each node corresponds to an identified salient feature and associated with information comprising a type and characteristics of the corresponding identified salient feature. The arcs correspond to a topological arrangement of the identified salient features. An optimal inexact structural match is determined between the nodes of the first graph and the nodes of the second graph so as to form matched sets of nodes comprising one node from each graph. A node lacking a determined match is matched with a null node. The characteristics of each of the nodes in a matched set are compared to one another to identify variances among the compared characteristics. A corresponding score is generated for each of the nodes in the matched set based on each of the identified variances. The nodes are identified whose appearances vary more than pre-specified thresholds based on the scores corresponding thereto, as well as the nodes matched with the null nodes.
These and other aspects, features and advantages of the present invention will become apparent from the following detailed description of preferred embodiments, which is to be read in connection with the accompanying drawings.
REFERENCES:
patent: 5133020 (1992-07-01), Giger et al.
patent: 5579360 (1996-11-01), Abdel-Mottaleb
patent: 5982915 (1999-11-01), Doi et al.
patent: 6032678 (2000-03-01), Rottem
patent: 6125194 (2000-09-01), Yeh et al.
patent: 6246782 (2001-06-01), Shapiro et al.
patent: 6282305 (2001-08-01), Huo et al.
Murthy Sreerama K.
Novak Carol L.
Qian Jianzhong
Wu Zhenyu
Johns Andrew W.
Nakhjavan Shervin
Paschburg Donald B.
Siemens Corporate Research Inc.
LandOfFree
Computer-assisted diagnosis method using correspondence... does not yet have a rating. At this time, there are no reviews or comments for this patent.
If you have personal experience with Computer-assisted diagnosis method using correspondence..., we encourage you to share that experience with our LandOfFree.com community. Your opinion is very important and Computer-assisted diagnosis method using correspondence... will most certainly appreciate the feedback.
Profile ID: LFUS-PAI-O-2821016