Image analysis – Applications – Biomedical applications
Reexamination Certificate
2001-05-15
2004-06-29
Patel, Jayanti K. (Department: 2625)
Image analysis
Applications
Biomedical applications
C378S004000, C378S019000, C378S062000
Reexamination Certificate
active
06757418
ABSTRACT:
BACKGROUND
1. Technical Field
The present invention generally relates to image composing and, in particular, to a method and system for automatic Computed Radiography (CR) image composition by white band detection and consistency rechecking.
2. Background Description
In a new type of X-ray image acquisition referred to as Computed Radiography (CR), cassettes containing storage phosphor plates are placed one after another so that anatomy larger than the size of the plates can be imaged by stitching individual images together. Since all the images are acquired at the same time, there is no content difference in the overlap region of neighboring image plates. This is not the case in auto-stepping, where the acquisition device moves stepwise after each acquisition of a small image. In general, when auto-stepping is employed, distortions exist in overlap regions due to depth differences.
FIG. 1
is a diagram illustrating a simplified view of a Computed Radiography (CR) image acquisition, according to the prior art.
Since individual images are digitized separately, there is a need to seamlessly combine the individual images after digitization. While such composition could be performed manually, manual approaches to composition are undesirably labor-intensive. Existing automatic methods that are based solely on cross correlation do not operate reliably, since the overlap between successive images is usually very small.
Accordingly, it would be desirable and highly advantageous to have a method and system for accurately and automatically combining Computed Radiography (CR) individual images after digitization.
SUMMARY OF THE INVENTION
The problems stated above, as well as other related problems of the prior art, are solved by the present invention, a method and system for automatic Computed Radiography (CR) image composition by white band detection and consistency rechecking.
The present invention provides a method and system for detecting the overlap region of successive images so that a mosaic of the images can obtained automatically and accurately.
According to an aspect of the invention, there is provided a method for automatic Computed Radiography (CR) image composition by white band detection and consistency rechecking. The method includes the step of receiving an original CR image pair for composition. For each image in the original CR image pair, the following steps are performed. A horizontal gradient image and a vertical gradient image are generated. The vertical gradient image is projected onto a y-axis of the each image to obtain a y-axis projection. Candidate edge positions are identified from among at least two white band edge positions according to a relative value of each of the at least two white band edge positions with respect to a percentage value of an absolute maximum value of the y-axis projection. An intensity change constancy of the candidate edge positions is determined to identify the candidate edge positions having orientation angles less than a pre-defined threshold angle with respect to a horizontal. The orientation angles are orthogonal with respect to angles of maximum intensity change for pixels on the candidate edge positions. An intensity change value is determined to verify intensity value differences on two sides of the candidate edge positions with respect to predefined criteria. An error function is defined to respectively obtain an error value for each of the candidate edge positions with respect to the intensity change constancy and the intensity change value. The candidate edge positions having the error value below a pre-specified threshold are selected. A cross-correlation score is computed for the selected candidate edge positions, by comparing a consistency of the selected candidate edge positions against image data corresponding to the original image pair. A final overlap is identified for the original image pair, based upon the cross-correlation score.
According to another aspect of the invention, I (x,y) represents the each image in the original CR image pair. I
x
(x,y) and I
y
(x,y) respectively represent the horizontal gradient image and the vertical gradient image. The y-axis projection is equal to
P
(
y
)
=
∑
x
I
y
(
x
,
y
)
.
According to yet another aspect of the invention, the absolute maximum value of the y-axis projection is equal to
Max_Py
=
max
y
&LeftBracketingBar;
P
(
y
)
&RightBracketingBar;
.
According to still yet another aspect of the invention, &OHgr;={c|c=1,2, . . . , N
c
} represents a set of indices for the candidate edge positions having y-coordinates at {y
c
|c=1,2, . . . , N
c
} in the each image. The angles of maximum intensity change for the pixels on a candidate edge position c from among the candidate edge positions are R
c
(x)=arctan(I
y
(x,y
c
),I
x
(x,y
c
); x&egr;D
c
,c&egr;&OHgr;, where D
c
is a set of points having a vertical gradient that is non-zero.
According to a further aspect of the invention, the set of points D
c
is used to exclude over-saturated and under-exposed pixels.
According to a still further aspect of the invention, an orthogonal angle to R
c
(x) is R
⊥
c
(x). An average angle of R
⊥
c
(x) represents an orientation angle of the candidate edge position c, and is equal to A
c
,
A
c
=
∑
x
∈
D
c
R
c
⊥
(
x
)
/
&LeftBracketingBar;
D
c
&RightBracketingBar;
.
|D
c
| is a number of elements in the set of points D
c
.
According to an additional aspect of the invention, the method further includes the step of excluding any of the candidate edge positions having an average angle below a pre-determined threshold from further consideration.
According to yet an additional aspect of the invention, the step of determining the intensity change constancy includes the following steps, performed for each of the candidate edge positions. A direction of maximum intensity change is determined for each of pixels on the each of the candidate edge positions. An orthogonal angle with respect to the direction of maximum intensity change is determined. A deviation is calculated of an orientation angle, measured as an average angle of the orthogonal angle, from the horizontal. The candidate edge positions having the deviation less than the pre-defined threshold are identified.
According to still yet an additional aspect of the invention, the step of determining the intensity change value includes the steps of measuring a relative intensity change across a candidate edge position c at two different, pre-defined offset values, and verifying that the relative intensity change satisfies the pre-defined criteria.
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: 4956859 (1990-09-01), Lanza et al.
patent: 5539838 (1996-07-01), Shimura
patent: 5557687 (1996-09-01), Hara
patent: 5901240 (1999-05-01), Luo et al.
patent: 6157698 (2000-12-01), Pietikainen et al.
patent: 6212291 (2001-04-01), Wang et al.
Qian Jianzhong
Schramm Helmut
Wei Guo-Qing
Wu Zhenyu
F. Chau & Associates LLP
Paschburg Donald B.
Patel Jayanti K.
Siemens Corporate Research Inc.
Tabatabai Abolfazl
LandOfFree
Method and system for automatic computed radiography (CR)... 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 and system for automatic computed radiography (CR)..., we encourage you to share that experience with our LandOfFree.com community. Your opinion is very important and Method and system for automatic computed radiography (CR)... will most certainly appreciate the feedback.
Profile ID: LFUS-PAI-O-3364174