Image analysis – Image transformation or preprocessing
Reexamination Certificate
2000-09-28
2004-09-07
Ahmed, Samir (Department: 2623)
Image analysis
Image transformation or preprocessing
Reexamination Certificate
active
06788827
ABSTRACT:
BACKGROUND OF THE INVENTION
The invention concerns an image processing method and system for following a moving object from one image to an other in an image sequence. The invention also concerns an X-ray examination medical apparatus using said system and method.
The invention finds applications in the industry of medical imaging.
SUMMARY OF THE INVENTION
An image Processing Method for Representing Moving Images is already known from the publication entitled “Representing Moving Images with layers”, in IEEE TRANSACTIONS ON IMAGE PROCESSING, VOL.3, No. 5, SEPTEMBER 1994, by John Y. A. Wang and Edward H. Adelson.
In order to avoid encoding images using high-level machine vision concepts such as 3-D object recognition, this publication proposes an image coding system which involves an image representation based on superimposed image layers. This publication further proposes to apply such a representation to the coding of video sequences. In digital systems, complex motions such as affine transformations including combinations of translation, rotation, dilation and shear, may be used. Such a representation involves an adequate description of the motions in a first sequence. For instance, motion of a hand waving on a moving background is studied. Then, given this first sequence, it is searched to invert the process by which the motions have been generated. To that end, the sequence is decomposed into a set of layers which may be compounded so as to generate said first sequence. Each layer comprises an intensity map, an opacity or transparency map referred to as alpha map and a velocity map. In addition the layers are assumed to be ordered in depth. The intensity map and alpha map are warped together. A delta map may also be allowed, for serving as an error signal to update the intensity map over time. Once a description in terms of layers is given, the image sequence may be straight forward generated. The difficult part is determining the layered representation given the input sequence. There, synthesis is easy, but analysis is hard. The representation is mono-unique. There will be many different descriptions that lead to the same synthesized image sequence. Thus, a layered representation can serve to specify the decoding process, while the encoding process remains open to further improvement by individual users.
In the cited publication, it is proposed, for solving the analysis problem, to simplify the representation in such ways as the alpha channel is binary, and there are no delta maps used. It is important to note from the cited publication, that-the method is assumed to be applied to stable objects undergoing smooth motions; that, in the images, which are binary, the objects are either completely transparent or completely opaque; that the velocity maps are restricted to affine transformations.
The solution proposed in the cited publication is not appropriate to medical image processing. A problem is that, in the field of medical imaging, some objects represented in images, for instance in X-ray images, do not move at all; some other objects periodically dilate so their walls have no smooth motion. Thus, they may not be defined as stable objects having smooth motions and they are not allowed to be segmented based on information related to their motion. An other problem is that the objects are translucent and, especially in an X-ray images, they are never completely transparent or completely black, but instead, several organs are seen through one another while superimposed. So, these organs of the image are not allowed to be represented by binary intensity values.
An aim of the invention is the improvement of medical images in order to make easier the visualization of objects in those images and further make easier the diagnosis of the practitioner based on said images. So, it is an aim of the invention to provide means for automatically following a given object of interest from one image to another in an image sequence. In medical applications, following automatically an organ, for instance defined by its contour, may be very useful for a physician in order to analyze the movements of said organ. But organ contours are often very fuzzy in the images so that following an organ location from one image to an other in the sequence may be difficult for the physician. For instance, it is very useful to follow the movements of the heart left ventricle in order to identify and localize abnormalities. But the heart left ventricle movements are especially difficult to follow because the heart left ventricle contour has no smooth movements and is situated in a cluttered environment formed by other organs having their own movements.
This aim is reached by a method according to Claim
1
. Such a method is favorably carried out by a system according to Claim
11
. Such a system may be favorably associated to an X-ray examination medical apparatus according to Claim
12
.
An advantage of the invention is that the method or system may be carried out successfully even in the case when the object of interest to be followed between two images of a sequence is a translucent object having a fuzzy contour represented in a cluttered environment. An other advantage is that the method is robust so that the object of interest may be followed automatically throughout a whole image sequence, basing the process on one image to go to the next image and so on from the last processed image to the next, until the end of the image sequence is reached. An other advantage of the invention is that it is not limited to be applied to an image sequence representing an object having smooth movements uniquely described by affine functions: the object may have any movements described by a parametrical function.
REFERENCES:
patent: 5293574 (1994-03-01), Roehm et al.
patent: 0609709 (1994-07-01), None
Automatic contour detection by encoding knowledge into active contour models, by Gerad et al, IEEE 1998.*
Gros P et al: “Using Local Planar Geometric Invariants to Match and Model Images of Line Segments” Computer Vision and Image Understanding, US, Academic Press, vol. 69, No. 2 Feb. 1, 1998, pp. 135-155.
“Representing Moving Images with Layers” by John Y.A. Wang and Edward H. Adelson in IEEE Transactions on Image Processing, vol. 3, No. 5, Sep., 1994, pp. 625-638.
Ahmed Samir
Bali Vikkram
Koninklijke Philips Electronics , N.V.
Vodopia John F.
LandOfFree
Image processing method and system for following a moving... 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 system for following a moving..., we encourage you to share that experience with our LandOfFree.com community. Your opinion is very important and Image processing method and system for following a moving... will most certainly appreciate the feedback.
Profile ID: LFUS-PAI-O-3215043