Image analysis – Applications – Biomedical applications
Reexamination Certificate
1993-12-28
2002-10-15
Chang, Jon (Department: 2623)
Image analysis
Applications
Biomedical applications
C382S274000, C128S922000
Reexamination Certificate
active
06466689
ABSTRACT:
BACKGROUND OF THE INVENTION
1. Field of the Invention
The present invention relates to a system and method for digital radiography, and more particularly to a system and method of digital chest radiography producing a single optimized image.
2. Discussion of the Background
Digital radiography provides the ability to manipulate a radiographic image in order to improve detection of abnormal findings. The adjustable parameters include density, contrast and edge enhancement by digital unsharp masking. The chest radiograph presents a particularly difficult challenge because of the large variation in x-ray transmission between the lungs and the mediastinum, as well as the great variety of abnormalities which may occur. These abnormalities range from relatively large “low frequency” types of findings such as nodules and air space infiltrates, to small “high frequency” findings such as pneumothorax and fine interstitial lung disease. It is therefore difficult to devise a single form of processing which is ideal for all applications.
Storage phosphor computed radiography (CR) is the most widely used digital technique in general radiography. This system was developed by Fuji Photo Film Company, and is extensively discussed in the patent literature. CR imaging systems are configured to produce hard copy consisting of both a mildly processed image, which resembles a conventional radiograph, and a more heavily processed version of the same image. The two images are printed on a single piece of film, typically 10 inches by 14 inches in size.
The image processing parameters of the imaging systems include the characteristic curve, the average contrast gradient contrast, unsharp mask filtering (UMF) and the processing curve. The characteristic curve is a measure of the optical density of the image, that is, how light or dark the image is, as a function of exposure or X-ray intensity, and provides contrast gradient information, while the processing curve represents the relative degree of UMF as a function of pixel value. UMF parameters include the mask size (typically 5.50 mm) and the weighting factor. The weighting factor, which governs the intensity of UMF, varies from 0.5 in the mildly processed standard image to 5.0 in the heavily processed standard image. A more detailed discussion of unsharp mask filtering and its relation to chest radiography are provided by Pratt, “Digital Image Processing”, John Wiley & Sons, New York (1978), MacMahon et al, “The Effect of Digital Unsharp Masking on the Detectability of Interstitial Infiltrates and Pneumothoraces”,
Proc. SPIE
, vol. 555, pp. 246-252 (1985), Ishida et al, “High quality digital radiographic images: Improved detection of low-contrast objects and preliminary clinical studies”,
Radiographics
, vol. 3, pp. 325-328 (1983), Loo et al, and “Investigation of basic imaging properties in digital radiography: 4. Effect of unsharp masking on the detectability of simple patterns”,
Med. Phys
. vol. 12, pp. 209-214 (1985).
In the mildly processed image, a characteristic curve as shown in
FIG. 1A
is typically used to retain good contrast in the lungs. The characteristic curve of
FIG. 1A
shows that at low pixel values, the optical density is slowly varying. As the pixel value increases, the optical density varies more rapidly. At higher pixel values, the optical density levels off. A typical chest image obtained using the characteristic curve of
FIG. 1A
is shown in FIG.
2
A. This moderately steep contrast gradient provides good contrast in the periphery of the lung in a standard dual image CR chest format, for example, but provides poor mediastinal detail. This approximates to a poor quality conventional screen/film radiograph, such as might be achieved with a moderately high contrast film. This is clearly not acceptable for use as a single image, as information in low density areas is deficient.
The more heavily processed standard image uses a characteristic curve as shown in FIG.
1
B. This more heavily processed image provides impressive detail in low density areas of the image such as the mediastinum and soft tissues. Fine details, such as septal lines and pleural fissures are also enhanced. A typical chest image obtained using the characteristic curve of
FIG. 1B
is shown in FIG.
2
B. The enhancement is achieved by using a relatively straight characteristic curve with a low overall contrast gradient, in combination with pronounced UMF. The flat characteristic curve produces a wide latitude effect, with retention of information in both low and high density areas. This would produce an unacceptably “flat” appearance in the absence of UMF. The processing effectively restores local contrast and enhances high frequency detail. However, the conspicuousness of pulmonary infiltrates and other relatively large low contrast abnormalities is markedly diminished in the more heavily processed image. This is due to the low overall contrast, as well as an increase in appearance noise due to UMF. Therefore, this type of image, if used in isolation, is also unacceptable. The ideal single image would retain large area contrast in the periphery of the lungs, while enhancing local contrast in low density areas to increase visibility of mediastinal, retromediastinal and retrodiaphragmatic details.
The prior art system uses a processing curve as shown in FIG.
3
. This curve shows the relationship between the intensity of the UMF applied to the image and the local pixel value of the image.
FIG. 3
gives the pixel value as a function of weighting factor fraction, which is the degree of the UMF blurred component added to the processed image, with 1.0 indicating that the degree of blurred component added is the same as the original digitized image. The UMF weighting factor is altered according to local image density. In this case, the processing curve is non-linear as different intensities of UMF are applied according to the density of the image. In the very low density areas, such as may occur in the mediastinum and upper abdomen, the UMF weighting factor is 0. The weighting factor rises rapidly to 1.0 (100%) in the higher density areas such as the lungs.
FIGS. 2A and 2B
considered together illustrate the standard dual image format where the mildly and heavily processed images are placed on a single film. While it is clear that these dual images are complimentary to some extent, it has not been demonstrated that this dual standard image format is optimal from either a diagnostic or operational point of view. Although the use of dual images is logical in theory, this format requires double the amount of film necessary for a given image size, and provides half the image size for a given area of film when both the mildly and heavily processed images are printed on a single film. Having two images printed on a single film produces images which are undesirably small, and this small size both impedes the effectiveness of the system and the diagnostic accuracy. The small size of the images has been found deficient by radiologists and clinicians.
SUMMARY OF THE INVENTION
It is therefore an object of the present invention to provide a novel system and method of digital radiography which overcomes the disadvantages of the prior art noted above.
It is another object of the present invention to provide a system and method of digital radiography which produces a single image which improves the visibility of detail in low density areas while maintaining good overall contrast.
These and other objects are achieved by a digital imaging system having means for obtaining a digital radiological image of a subject having at least one low density region, image processing means for processing said digital image using weighting factors, said processing means using first weighting factors in said at least one low density region of said digital image, and second weighting factors smaller than said first weighting factors for regions of said digital image other than said at least one low density region, storage means for storing said digital image and said processed digital image, and
Arch Development Corp.
Chang Jon
LandOfFree
Method and system for digital radiography 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 digital radiography, we encourage you to share that experience with our LandOfFree.com community. Your opinion is very important and Method and system for digital radiography will most certainly appreciate the feedback.
Profile ID: LFUS-PAI-O-2965132