Image analysis – Applications – Biomedical applications
Reexamination Certificate
2000-08-01
2004-05-18
Patel, Jayanti K. (Department: 2625)
Image analysis
Applications
Biomedical applications
C382S298000, C382S300000, C345S608000, C345S609000, C358S525000
Reexamination Certificate
active
06738498
ABSTRACT:
STATEMENT REGARDING FEDERALLY SPONSORED RESEARCH & DEVELOPMENT (IF APPLICABLE)
Not applicable.
BACKGROUND OF THE INVENTION
The preferred embodiments of the present invention generally relate to digital image magnification, and in particular relate to a method and apparatus for tissue dependent filtering for digital image magnification.
Doctors and technicians have at their disposal a wide range of ultrasound, x-ray, nuclear, and other medical diagnostic imaging systems with which to examine patients. The capabilities of these medical diagnostic imaging systems have increased dramatically since their introduction. Spurred on by the development of inexpensive but very sophisticated, powerful, and fast processing circuitry, designers of medical diagnostic imaging systems continue to add and enhance a wide range of device functions for medical diagnostic imaging systems. Thus, for example, an x-ray imaging system may include magnification, 2D or 3D imaging, Doppler overlay, Colorflow scans, image frame recording and playback capabilities, image annotation and archiving, panning, and the like.
Medical diagnostic imaging systems, such as x-ray imaging systems, have long been accepted medical diagnostic tools. For instance, x-ray imaging systems are commonly used to capture, as examples, thoracic, cervical, spinal, cranial, and abdominal images that often include information necessary for a doctor to make an accurate diagnosis. X-ray imaging systems typically include an x-ray source and an x-ray sensor. When having a thoracic x-ray image taken, for example, a patient stands with his or her chest against the x-ray sensor as an x-ray technologist positions the x-ray sensor and the x-ray source at an appropriate height. X-rays produced by the source travel through the patient's chest, and the x-ray sensor then detects the x-ray energy generated by the source and attenuated to various degrees by different parts of the body. An associated control system obtains the detected x-ray energy from the x-ray sensor and prepares a corresponding diagnostic image on a display.
The x-ray sensor may be a conventional screen/film configuration, in which the screen converts the x-rays to light that exposes the film. The x-ray sensor may also be a solid state digital image detector. Digital detectors afford a significantly greater dynamic range than conventional screen/film configurations, typically as much as two to three times greater.
Medical diagnostic images may be used for many purposes. For instance, internal defects in a target object may be detected. Additionally, changes in internal structure or alignment may be determined. Furthermore, the image may show the presence or absence of objects in the target.
For a variety of purposes, it may be desirable to magnify at least portions of medical diagnostic images. Slight defects or small objects in a patient's body may be examined more closely in a magnified digital image. Magnification of a digital image may be achieved by modifying pixels or picture elements that comprise the digital image. Several techniques may be used to magnify a digital image, such as pixel replication and interpolation. Pixel replication simply copies the pixels of a digital image to enlarge a digital image without adjusting the resolution of the digital image. Pixel replication results in a low quality magnified digital image. Interpolation resizes the image and adjusts the resolution based on individual pixel data values and relationships between the pixel data values. Using interpolation, each pixel from the original digital image is divided into multiple new pixels. The data values assigned to the new pixels are based upon relationships between surrounding original pixel data values. Methods of interpolation include nearest neighbor interpolation and bilinear interpolation.
Nearest neighbor interpolation is a fast technique and uses relationships between the new pixel and an original adjacent pixel. However, nearest neighbor interpolation may produce a “stair stepped” effect around diagonal lines and curves in the digital image. Bilinear interpolation uses four original pixels. Bilinear interpolation may produce a smoother, more accurate image than nearest neighbor. However, while bilinear interpolation may smooth the edges of objects in the digital image, it may obscure some extreme data values in the digital image. Some high quality magnification methods, such as bicubic interpolation, are more accurate magnification methods. However, high quality magnification methods, such as bicubic interpolation, have not heretofore been included in graphic software that is useful with medical diagnostic imaging systems. Bicubic interpolation calculates a new pixel value using sixteen original pixel values. Bicubic interpolation is typically very computationally expensive.
There is a need to provide high quality techniques for digital image magnification for multiple medical diagnostic imaging systems. There is also a need for high quality magnification methods that adapt to the content of the digital images to be magnified. There is also a need for high quality magnification methods that may be easily and repeatedly used for a variety of images.
Thus, a need exists for a method and apparatus for tissue dependent filtering for digital image magnification.
BRIEF SUMMARY OF THE INVENTION
A preferred embodiment of the present invention provides a method and apparatus for tissue dependent filtering for digital image magnification. The method and apparatus alter the spatial characteristics of digital images to magnify at least a portion of a digital image. A library is created with sets of representative images for particular anatomies using particular imaging modalities. Two-dimensional convolution filter coefficients are estimated using the sets of representative images. Using a set of representative images for a particular anatomy, a function is formed which represents the mean square difference between pixel values obtained using bicubic interpolation and pixel values obtained using a preferred embodiment. The function is then minimized using a simulated annealing method. The coefficients in the global minimum of the function constitute the convolution filter coefficients for the particular anatomy at a particular magnification value.
The method and apparatus use bilinear interpolation and convolution filters to approximate bicubic interpolation for digital image magnification. First, the method and apparatus apply bilinear interpolation to the digital image. Then, a set of representative convolution filter coefficients is selected for the digital image. A two-dimensional convolution filter is applied to the digital image. The two-dimensional convolution filter is based on the convolution filter coefficients calculated from the set of representative images.
REFERENCES:
patent: 5235432 (1993-08-01), Creedon et al.
patent: 5847714 (1998-12-01), Naqvi et al.
patent: 5917961 (1999-06-01), Huonder
patent: 6061477 (2000-05-01), Lohmeyer et al.
patent: 6101234 (2000-08-01), Ali et al.
patent: 6101235 (2000-08-01), Zavaljevski et al.
patent: 6236766 (2001-05-01), Zavaljevski et al.
patent: 6501818 (2002-12-01), Ali et al.
patent: 6542242 (2003-04-01), Yost et al.
Carter Aaron
Dellapenna Michael A.
GE Medical Systems Global Technology Company LLC
McAndrews Held & Malloy Ltd.
Patel Jayanti K.
LandOfFree
Method and apparatus for tissue dependent filtering for... 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 tissue dependent filtering for..., we encourage you to share that experience with our LandOfFree.com community. Your opinion is very important and Method and apparatus for tissue dependent filtering for... will most certainly appreciate the feedback.
Profile ID: LFUS-PAI-O-3233019