Radiant energy – Photocells; circuits and apparatus – Photocell controlled circuit
Reexamination Certificate
1998-10-09
2001-04-24
Le, Que T. (Department: 2878)
Radiant energy
Photocells; circuits and apparatus
Photocell controlled circuit
C382S279000
Reexamination Certificate
active
06222173
ABSTRACT:
FIELD OF THE INVENTION
The present invention relates to digital image processing. The invention more specifically relates to sharpening and re-sampling of medical images.
STATE OF THE ART
In several medical imaging techniques, a visible image of an object under examination is generated by means of thermal reproduction techniques such as dry thermal printing.
The sharpness of the reproduced images, especially when they are reproduced on a transparent thermal sensitive medium is not always adequate for medical diagnosis. The images are considered un-sharp or not crispy.
The degree of un-sharpness experienced by the viewer, e.g. the physician or the radiologist may depend on the kind of image that has been reproduced, on the perception of the viewer, on the characteristics of the image acquisition etc.
The un-sharpness (also referred to as smoothing) may be caused by thermal effects that have occurred in the thermal reproduction device such as heat distribution among the thermal elements of the thermal head part of the printer. It may also be caused by slow temperature increase and decrease of the heating elements in the thermal head. Although compensation for these effects is often provided, it may still not eliminate all non-uniformities.
Still other causes of un-sharpness may occur.
Furthermore, when images acquired by a digital image acquisition apparatus are reproduced, there is often a need for re-scaling of the pixel matrix. Images may need to be re-scaled for example because the pixel matrix of the acquired digital image does not match with the pixel matrix provided by a reproducing apparatus by means of which a hard copy of the digital image will be produced. Also lay-out requirements such as reproduction of a multiplicity of images on a single film (often applied in tomography), may demand for image re-scaling.
The above problems, i.e. sharpening and re-scaling, have been addressed separately in the prior art.
Sharpening techniques are known according to which (a) digital spatial convolution filter(s) is(are) applied to the digital image. These digital spatial convolution filters used for image sharpening do not provide re-sampling nor image scaling.
An example of such a technique is the unsharp masking technique wherein smoothing low pass filters such as mean filters, Gaussian filters etc. are used.
Another example consists of the application of a highpass filter such as the Laplacian or Gaussian filter.
In the prior art also image processing techniques are known which are based on a digital spatial interpolating convolution to be applied to images, e.g. scaled image, for creating intermediate pixels in between the pixels of the original pixel matrix of an image.
An example of such a technique is referred to as high resolution cubic spline interpolation.
However, images often need to be re-scaled in addition to being sharpened.
Application of the prior art methods requires both problems to be tackled separately, i.e. one after the other. The prior art solutions hence are computationally expensive, they demand a lot of processing time and are hence not adapted for use in medical imaging where a physician wants to dispose of a hard copy image as soon as possible after examination in order to make a diagnosis.
OBJECTS OF THE INVENTION
It is thus an object of the present invention to provide a method of sharpening and re-scaling an image that can be performed fast and that does not demand a lot of processing time.
It is a further object to provide such a method that can easily be adapted to different kinds of images as well as to the user's taste.
It is still a further object to provide such a method of sharpening and re-scaling that is adapted to the requirements of medical imaging.
Further objects of the invention will become apparent from the description given below.
SUMMARY OF THE INVENTION
The objects of the present invention are achieved by a method according to claim
1
.
The method of the present invention is applicable to images provided by means of a digital signal representation.
Although the method was specifically developed for application to medical images, it can be applied to all kinds of images that require image sharpening and for which varying image scaling or re-sampling is desired.
The method can be implemented by software as well as by a mixed hardware-software embodiment.
The method provides a simple implementation of a combined image sharpening and image re-sampling technique. In medical environments such as hospitals, physicians want to be able to make a diagnosis on a hard copy image as soon as possible after an examination. Care is thus taken to use a fast hard copy system such as the dry imaging system and to use a fast working image processing system such as the processing of the present invention which is computationally less expensive than the prior art techniques. It provides enhanced images in fast and accurate way and is thus very well suited for use in a medical environment.
The method of the present invention does not introduce unwanted artifacts e.g. due to aliasing, overshoots, contouring etc. Consequently it does not have any negative influence on the diagnosis.
The method of the invention makes use of either a one or a two-dimensional kernel.
When a two-dimensional kernel is used, the image can be filtered in a single pass, whereas the use of one-dimensional kernels implies two passes, one for a first, e.g. horizontal filtering and one for a second, e.g. vertical filtering of a two-dimensional image.
In a preferred embodiment (down-loadable) convolution tables are used so that the image processing characteristics can be easily adapted to different kinds of images or to the taste of the user.
According to the invention the kernel elements are obtained by selecting a first set of kernel values so that a convolution of the image by means of a kernel with said first set of kernel values generates a sharpened image and interpolating between kernel values of the first set.
The interpolation preferably is based on a third order function. Preferably a cubic spline interpolation between the values of the first kernel is used such that a desired interpolation function is approximated. The desired interpolation function is preferably a sin ×/× function.
In order to be able to obtain said approximation, the following constraints must be applied:
in case of a scaling factor equal to 1/1 the sharpening kernel values need to be used (values of the first kernel),
the interpolation function must be symmetric around the center point of the kernel,
the slope in the center point is zero,
the function is continuous in all points,
the slope of a spline in the end point of a spline should be equal to the slope in the start point of an adjacent spline.
Other interpolation techniques can be use such as linear interpolation (with triangular function) etc. However, cubic spline approximation is preferred because it does not introduce undesired effects which could have a negative influence on the diagnosis.
REFERENCES:
patent: 4969043 (1990-11-01), Pothier
patent: 5008752 (1991-04-01), Van Nostrand
patent: 5168375 (1992-12-01), Reisch et al.
patent: 5347590 (1994-09-01), Nonnweiler et al.
patent: 5410616 (1995-04-01), Kidd
patent: 5627953 (1997-05-01), Yen
patent: 5666391 (1997-09-01), Ohnesorge et al.
patent: 0396517 (1990-07-01), None
patent: 9000780 (1990-01-01), None
Meeussen Dirk
Tytgat Bart
AGFA-GEVAERT
Baker & Botts L.L.P.
Le Que T.
LandOfFree
Image sharpening and re-sampling method 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 sharpening and re-sampling method, we encourage you to share that experience with our LandOfFree.com community. Your opinion is very important and Image sharpening and re-sampling method will most certainly appreciate the feedback.
Profile ID: LFUS-PAI-O-2468254