Image analysis – Applications – Biomedical applications
Reexamination Certificate
2000-08-11
2004-08-17
Boudreau, Leo (Department: 2621)
Image analysis
Applications
Biomedical applications
C382S254000
Reexamination Certificate
active
06778692
ABSTRACT:
BACKGROUND OF THE INVENTION
The present invention relates to the art of image processing, and more particularly to image enhancement, image smoothing, and other image improvement techniques for image processing. Medical diagnostic images have commonly been subject to image degradation from noise, system imperfections, and the like. Various image processing techniques have been utilized to remove the effects of the noise. For example, one technique adjusts each pixel in accordance with the mean of surrounding pixels and the variance or difference between the surrounding pixels. Each filter enhanced pixel value g′(i,j) is a weighted average of the local mean and variance values:
g
′(
i,j
)=&ggr;(
i,j
)+
k[g
(
i,j
)−&ggr;(
i,j
)]
where g(i,j) is the original unprocessed image, &ggr;(i,j) is a local mean, g(i,j)−&ggr;(i,j) is the variance, and k is a constant that weights the relative contributions therebetween. According to this technique, when k is set larger than 1, the variance or difference between the local mean value, hence the contribution of the measured gray scale level of the pixel g(i,j) is magnified. As k is set smaller, the image is smoothed or blurred as if acted upon by a low-pass filter. At the extreme at which k is set equal to zero, each pixel value is replaced by the local mean of the neighboring pixel values.
One of the drawbacks in this technique is the difficulty of selecting an appropriate value for the weighting factor k. The smaller k is set, the more the processed image is blurred and the more difficult it becomes to withdraw accurate diagnostic information from the processed image. As k is set larger, edges and fine details become enhanced. However, noise becomes enhanced at the same time. Frequently, in a medical image, the selected weighting factor k is too large for some regions and too small for other regions.
Other techniques involve setting the k for each pixel equal to the square root of the ratio of a preselected desirable local variance to the actual local variance of the selected pixel. Although this type of variable weighting factor can achieve better resultant images than the constant weighting factor, there is still room for improvement. In accordance with the present invention, an improved method for determining weighting coefficients for pixel intensity values is provided.
BRIEF SUMMARY OF THE INVENTION
A method of improving images is provided. A body of image data is collected and signal to noise (SNR) values thereof are determined. The collected data is converted into an electronic image representation X which comprises a first array of pixel intensity values. Further, a low pass filtered image representation F comprising a second array of pixel intensity values is obtained. An improved electronic image representation P comprising a third array of pixel intensity values is generated by replacing each pixel intensity value of electronic image representation X by a linear combination of the replaced pixel intensity value from image X and the corresponding pixel intensity value from image F. An array of linear combination coefficients, also referred to herein as weighting coefficients, &agr;, (1−&agr;) determine the relative contributions of the replaced X pixel intensity value and the value of the corresponding pixel intensity from image F, respectively. The values of linear combination coefficients array &agr; are determined with respect to statistical image quality metrics, in particular with respect to signal-to-noise (SNR) ratio. In one embodiment of the invention weighting coefficients are pre-determined based on known theoretical SNR for particular imaging apparatus and target SNR values (e.g., the Poisson distribution for x-ray images and Gaussian for MRI images). The predetermined coefficients are stored in a look up table. In another embodiment of the invention, the array of weighting coefficients &agr; and (1−&agr;) are computed based on the measured SNR of the replaced pixel from image X and target SNR values so as to achieve a desired target SNR for image P. Target SNR values are determined by experimentation and human visual perception studies.
REFERENCES:
patent: 4590582 (1986-05-01), Umemura
patent: 4761819 (1988-08-01), Denison et al.
patent: 5051902 (1991-09-01), Hishinuma
patent: 5173776 (1992-12-01), Rigg et al.
patent: 5684720 (1997-11-01), Hein
patent: 5901241 (1999-05-01), Koljonen et al.
patent: 5911012 (1999-06-01), Bernard et al.
patent: 6211515 (2001-04-01), Chen et al.
patent: 6252931 (2001-06-01), Aach et al.
Boudreau Leo
General Electric Company
Miller Ryan J.
Patnode Patrick K.
Testa Jean K.
LandOfFree
Image processing method and apparatus including image... does not yet have a rating. At this time, there are no reviews or comments for this patent.
If you have personal experience with Image processing method and apparatus including image..., we encourage you to share that experience with our LandOfFree.com community. Your opinion is very important and Image processing method and apparatus including image... will most certainly appreciate the feedback.
Profile ID: LFUS-PAI-O-3359213