Image analysis – Image enhancement or restoration – Object boundary expansion or contraction
Reexamination Certificate
1999-09-28
2003-11-18
Johns, Andrew W. (Department: 2621)
Image analysis
Image enhancement or restoration
Object boundary expansion or contraction
C382S304000, C382S308000
Reexamination Certificate
active
06650788
ABSTRACT:
FIELD OF THE INVENTION
The present invention relates to the field of image processing systems, and more particularly to a digital system and method for binary morphological image processing.
BACKGROUND OF THE INVENTION
Conventional image processing systems are used in a variety of applications, including automatic object recognition systems which allow computers to identify particular objects or patterns in an image, as well as systems which enhance an image to make particular objects or patterns in the image more easily recognizable. Such systems typically operate by acquiring and digitizing an image into a matrix of pixels.
In non-grayscale, or binary, images each pixel is encoded in a single bit which is set to zero if the pixel is dark, or set to one if the pixel is illuminated. The image processor scans the digital image and processes the digital information to interpret the image.
One class of image processing systems is based on mathematical morphology, the science of processing and analyzing digital images through mathematical set transformations. Binary morphological processing was first investigated in the 1970s. See G. Matheron,
Random Sets and Integral Geometry
, Wiley, N.Y., 1975. Thereafter, the process was extended to grayscale morphology. See S. R. Sternberg, “Grayscale Morphology”,
Computer Vision, Graphics, and Image Processing
, vol. 35, pp. 333-355, 1986.
Morphological image processing techniques can be expressed in terms of two basic transformation operations: dilation and erosion. The dilation operation propagates pixel states representative of relatively high intensities throughout some region of an image, while the erosion operation contracts regions of relatively high intensity in an image.
Morphological image processing operates by using the geometric information of an image which is exactly specified in an image probe, also called a structuring element. Using this known geometrical information, particular features in the image can be extracted, suppressed, or preserved by applying morphological operators. Structuring elements can be constructed in one or more dimensions. Examples of common three-dimensional (3D) structuring elements include spheres, cones, bars, cylinders, parabaloids, and polyhedrons.
However, morphological dilation and erosion operations suffer from a significant drawback. Specifically, they can be very computationally time-consuming depending on the size of the structuring element used. See, for example, Cytocomputer described in A. P. Reeves, “Survey: Parallel Computer Architectures for Image Processing”,
Computer Vision, Graphics, and Image Processing
, vol. 25, pp. 68-88, 1984; and the Diff
3
Analyzer described in M. D. Graham and P. E. Norgren, “The Diff3 Analyzer: a Parallel/Serial Golay Image Processor”,
Real-Time Medical Image Processing
, M. Onoe et al. eds., pp.163, Plenum Press, New York, 1980. Other examples include the algorithms for decomposing a large two-dimensional (2D) structuring element described in, J. Xu, “Decomposition of Convex Polygonal Morphological Structuring Elements into Neighborhood Subsets”,
IEEE Trans, Pattern Analysis and Machine Intelligence
, vol. 13, no. 2, 153-162, Feb. 1991.
SUMMARY OF THE INVENTION
A binary morphological image processor which performs morphological erosion operations on an input image of any dimension by using a one-dimensional (1D) operator that constructs a structuring element of any size and shape. Using translation invariance, the input image and structuring element are first decomposed until they are in the 1D domain. The 1D structuring element is then spatially decomposed into smaller segments which operate in parallel to transform the input image. Structuring elements of any size and shape can be composed to perform the parallel erosion operations on an image of any dimension. Processing multiple 1D operations in parallel using the present invention significantly reduces processing time as compared to conventional systems.
REFERENCES:
patent: 4301443 (1981-11-01), Sternberg et al.
patent: 4395699 (1983-07-01), Sternberg
patent: 6192160 (2001-02-01), Sunwoo et al.
LandOfFree
Expandable binary morphological erosion operator does not yet have a rating. At this time, there are no reviews or comments for this patent.
If you have personal experience with Expandable binary morphological erosion operator, we encourage you to share that experience with our LandOfFree.com community. Your opinion is very important and Expandable binary morphological erosion operator will most certainly appreciate the feedback.
Profile ID: LFUS-PAI-O-3141206