Image analysis – Pattern recognition – Template matching
Reexamination Certificate
2000-03-24
2004-12-07
Dastouri, Mehrdad (Department: 2623)
Image analysis
Pattern recognition
Template matching
C382S280000, C382S131000, C382S132000
Reexamination Certificate
active
06829389
ABSTRACT:
BACKGROUND OF THE INVENTION
The present invention relates to the technology of image evaluation for checking images for any abnormalities such as unevenness. In particular, the invention relates to an image evaluating method and an image evaluating apparatus for assessing unevenness in images such as radiation images and those produced with digital radiographic apparatuses.
Various diagnostic apparatuses have heretofore been used in the medical field employing X-rays and other radiations. In radiology, X-rays passing through a subject are detected and an image is generated for diagnostic purposes. Images taken by simple chest radiography have been used widely for quite many years.
Apparatuses relying upon CR (computerized radiography), CT (computerized tomography) and MRI (magnetic resonance imaging) are in commercial use. The images generated by these apparatuses are either represented on CRT (cathode ray tube) displays or output on films as by LP (laser printers) for subsequent use in medical working fields to diagnose diseases.
Radiographic and other medical diagnostic apparatus are making progressive shift toward digitizing. By the word “digitizing” is meant a process in which the X-ray signal passing through a subject is converted to a digital signal, which is appropriately processed to generate an image suitable for diagnosis.
In the apparatus for CR, there is used a stimulable phosphor which, upon illumination with a radiation, stores part of its energy and upon illumination with exciting light such as visible light or laser light, produces stimulated light emission in accordance with the stored radiation energy. The radiation image information about a subject such as the human body is first recorded in a sheet of stimulable phosphor (commonly called “a stimulable phosphor sheet”), which is then scanned with exciting light such as laser light to produce stimulated light which is read photoelectrically as an electric signal, producing a digital picture signal. The radiation image recording and reading system used in CR is also called a digital radiographic apparatus.
The digital radiographic apparatus such as the CR apparatus used for medical diagnosis have been required to produce images of extremely high quality since unevenness and other image abnormalities can be a cause of wrong diagnosis.
The causes of image abnormalities are diverse and they may originate in the apparatus itself or a certain material such as the stimulable phosphor sheet. Evaluation is sometimes performed at the critical point beyond which an image abnormality is considered to affect the reliability in diagnosis. This requires an extremely subtle distinction and has been impossible to accomplish with machines using objective and quantitative criteria. As an alternative, unevenness and other image abnormalities have so far been evaluated by human eyes using threshold samples.
However, this visual approach involves individual differences, as well as inconsistency in reproducibility. The use of comparative samples does not ensure consistent image evaluation unless the operator is highly skilled. In other words, efficiency has been very low in the conventional visual assessment.
SUMMARY OF THE INVENTION
The present invention has been accomplished under these circumstances and has as an object providing a method for image evaluation that is capable of efficient objective assessment on the basis of a quantitative criterion in checking for image abnormalities.
Another object of the invention is to provide an apparatus for implementing the method.
The first object of the invention can be attained by an image evaluation method that checks for image abnormalities, comprising the steps of preliminarily setting a Mahalanobis space using a specified characteristic quantity extracted from specified image data, calculating a Mahalanobis distance using the Mahalanobis space and image data read from an image to be evaluated, and comparing the calculated Mahalanobis distance with a specified threshold value to check for the image abnormalities in the image to be evaluated.
Preferably, respective Mahalanobis distances in a main and an auxiliary scanning direction calculated from characteristic quantities for the image data about the image to be evaluated in the main and the auxiliary scanning direction, respectively and the respective Mahalanobis distances are compared with respective specified threshold values to check for the image abnormalities in the image to be evaluated.
Preferably, the specified characteristic quantity for forming the Mahalanobis space contains at least one of the following quantities of respective lines in the specified image data in a main or an auxiliary scanning direction, a sum, an average, a standard deviation, a Fourier transformed maximum power, a spatial frequency of the Fourier transformed maximum power, a maximum value, a minimum value and an amplitude between the maximum and minimum values.
Preferably, the specified image data and the image data read from the image to be evaluated are subjected to logarithmic conversion.
Preferably, the specified image data is read from at least one normal image which has not the image abnormalities at all.
The second object of the invention can be attained by an apparatus for evaluating images, comprising a Mahalanobis space setting unit which calculates preliminarily a specified characteristic quantity from specified image data and sets preliminarily a Mahalanobis space using the specified characteristic quantity, a Mahalanobis distance calculating unit which calculates a Mahalanobis distance using the Mahalanobis space and image data read from an image to be evaluated, and a comparing unit which compares the calculated Mahalanobis distance with a specified threshold value and thereby checks for image abnormalities in the image to be evaluated to evaluate the image.
Preferably, the specified image data is read from at least one normal image which has not the image abnormalities at all.
It is preferable that the apparatus further comprises an image reading unit which reads the specified image data and the image data from the image to be evaluated.
It is also preferable that the apparatus further comprises an extracting unit which extracts image samples from the specified image data and the image data from the image to be evaluated image data read by the image reading unit.
REFERENCES:
patent: 5003979 (1991-04-01), Merickel et al.
patent: 5710833 (1998-01-01), Moghaddam et al.
patent: 6117601 (2000-09-01), Kanazawa et al.
patent: 6243492 (2001-06-01), Kamei
patent: 6421553 (2002-07-01), Costa et al.
patent: 6427082 (2002-07-01), Nordstrom et al.
“An Introduction to the Quality Engineering Calculation Method” pp. 271-287, 1998.
Genichi Taguchi et al., “Robust Engineering” chapter 19, McGraw-Hill, pp. 222-224.
Arakawa Satoshi
Seto Noboru
Watanabe Kazuya
Dastouri Mehrdad
Fuji Photo Film Co. , Ltd.
Sughrue & Mion, PLLC
LandOfFree
Method and apparatus for evaluating 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 Method and apparatus for evaluating image, we encourage you to share that experience with our LandOfFree.com community. Your opinion is very important and Method and apparatus for evaluating image will most certainly appreciate the feedback.
Profile ID: LFUS-PAI-O-3309310