Image analysis – Image enhancement or restoration – Intensity – brightness – contrast – or shading correction
Reexamination Certificate
2000-03-10
2004-07-20
Boudreau, Leo (Department: 2621)
Image analysis
Image enhancement or restoration
Intensity, brightness, contrast, or shading correction
C382S132000, C382S131000
Reexamination Certificate
active
06766064
ABSTRACT:
BACKGROUND OF THE INVENTION
The present invention relates to a method, a memory medium and an apparatus implementing and performing a contrast-based dynamic range management (“C-DRM”) algorithm, for compressing an intensity dynamic range of an input image to a reduced intensity dynamic range supported by an available display device and, moreover, for doing so in a fashion which maximizes the displayed image contrast and detail. The compression is performed, in accordance with the invention, by directly and separately managing the mean (low frequency) and contrast (high frequency) content of the input image. This affords more deterministic behavior and reduced complexity, while enabling automatic adaptation to the dynamic range of the image, and results in minimizing the extent of compression required to enable displaying the image with optimized contrast on an available display device. The invention more particularly relates to such a method, a memory medium and an apparatus applicable to cardiac x-ray imaging and accordingly operable at video rates (typically 30 frames per second) and having low latency (i.e., a time duration from image acquisition to display of around 150 ms and thus effectively producing a real time display) to enable eye-hand coordination and which does not require interactive tuning of performance by a physician, permitting the physician to concentrate on performing an on-going medical procedure in reliance upon the displayed images, e.g., guiding a catheter through a blood vessel of a patient, without being distracted by any requirement for interactive tuning of the image acquisition, processing and display apparatus.
Digital X-ray imaging is a well-known, non-contact technique for observing, in real time, interior aspects of an object. In practice, an X-ray beam is generated and targeted on the object of interest. A detecting device is positioned on the other side of the object and detects the X-rays transmitted through the object. The detected X-ray signals are converted to digital signals that represent various features in the object, are further processed, and the resulting signal is displayed on a display device such as a CRT.
One of the fundamental image processing problems in digital X-ray imaging is the need to transform the intensity dynamic range of the input image to the dynamic range supported by an available display device. Typically, the intensity dynamic range of an image exceeds the dynamic range of the display mechanism by several times. The objective of the transformation accordingly is to compress the “DC” or mean component of the different regions comprising the image so that the dynamic range (typically 256 gray levels) of an available display device may be utilized in a fashion which maximizes the displayed image contrast—taking into account, as well, the generally recognized limitation of the human eye of discerning only 256 gray levels.
Common approaches to achieving such transformations are known as “unsharp masking” and “extended dynamic range (EDR)”, the latter a special adaptation of the former. The conventional approach to performing the EDR algorithm, in general, is simply to subtract a portion of the mean from the input signal. However, in some situations, this approach results in important contrast (higher frequency) information either being removed from the image or being artificially enhanced and thereby introducing artifacts. In certain improved implementations of the EDR algorithm, however, compression is less likely to result in a loss of higher frequency contrast information or an introduction of artifacts.
A more recent type of EDR function, employed in a cardiac feasibility study relating to a predecessor system relative to the present invention, is computed and implemented in the following manner:
y
(
i,j
)=GAMMA[
x
(
i,j
)−BOOST[
{overscore (x)}
(
i,j
)]], (1)
where:
y(i,j) is the (i,j)
th
pixel value of the output image;
x(i,j) is the (i,j)
th
pixel value of the input image; and
{overscore (x)}(i, j), i.e., (x_bar(i,j)), is the local spatial mean intensity value of the (i,j)
th
pixel, derived from a BOXCAR average. (In practice, intensity is directly related to X-ray count, but the relationship is rather complex. The x-rays are converted to photons which, in turn, are converted the electrons, in an x-ray imager. The Electrons are then digitized by an analog-to-digital converter (“A/D” or “ADC”) and basic image corrections are performed, such as adjustments to gain, offset and scaling, after which the image is ready to be processed for display.)
A graphical representation of the EDR processing is shown in FIG.
1
. The intensity value x of an input pixel (i,j) of an image is first processed by a BOXCAR (moving average) function
12
that determines the local mean intensity value at that (i,j) pixel location. (An “x” is used herein to designate an input intensity at a pixel location and, thus constitutes an individual scaler value; by contrast, an “X” designates the intensity image value, and thus is a vector value.) The BOXCAR function
12
utilizes a neighborhood of pixels, which includes and is centered on the input pixel, to calculate the local spatial mean intensity value {overscore (x)}(i,j)—(see, terms of Equation (1), supra).
As illustrated in
FIG. 1
, BOOST LUT
14
comprises a look-up table (LUT) which specifies the intensity reduction of the input image signal x(i,j) as a function of the local spatial mean intensity value {overscore (x)}(i,j). An adder (ADD)
18
combines the (negative) output of BOOST LUT
14
and the (positive) output of LUT
16
(see Equation (1)) and supplies the result to GAMMA LUT
20
, which then compresses the result of the unsharp masking, or subtraction, operation of an ADD
18
to 256 levels (8 bits per pixel, or 8 bpp). The LUTs
14
and
20
are indexed by the appropriate pixel intensity values given in equation (1). Thus, each of the BOOST LUT
14
and the GAMMA LUT
20
jointly manages both mean and contrast modification functions. Thus, this more current EDR processing algorithm, while an improvement over previous compression/transformation algorithms (e.g., which merely subtracted the local mean intensity signal from the input signal), is relatively complex, yet does not permit simultaneous, independent control of the mean and contrast modification functions.
Another problem with the current EDR processing algorithm is that of inconsistent contrast management resulting in exaggeration of negative contrast regions in the image. When a region in the image, such as a vessel filled with dye, for example, has an image intensity which is less than the surrounding local mean intensity value, EDR processing may exaggerate the negative contrast associated with the darker region when it subtracts the local mean intensity values from the intensity values associated with the darker region. This exaggerated negative contrast may result in artifacts, which can lead to misdiagnosis.
Where an image includes multiple areas having potentially differing mean levels of gray, or when images of objects embedded in such areas have respective, different gray levels, or if respective gray levels of an object and its background have similar values, the contrast parameters of the display window must be adjusted to enhance the visibility of these differences in order to obtain a diagnosis of the underlying structure being imaged. Thus, when a viewer's attention is shifted from one object to another, where the contrast of one combination of object and background differs significantly from the contrast of another combination of object and background, various display window parameters relating to contrast adjustment must be changed. Without such adjustments, the image will appear either excessively faint or excessively bright, such that all detail critical to an effective diagnosis is absent. As a result, in order to obtain a diagnosis, it is often necessary, during the course of shifting attention among areas of differing contrast, for the physic
Hopple Michael Robert
Langan David Allen
Lienard Jean
Nevin Robert Leland
Boudreau Leo
Dang Duy M.
General Electric Company
Patnode Patrick K.
Testa Jean K.
LandOfFree
Method and apparatus for performing a contrast based dynamic... 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 apparatus for performing a contrast based dynamic..., we encourage you to share that experience with our LandOfFree.com community. Your opinion is very important and Method and apparatus for performing a contrast based dynamic... will most certainly appreciate the feedback.
Profile ID: LFUS-PAI-O-3240433