## Efficient 2-D Grayscale Morphological Transformations With Arbitrary Flat Structuring Elements

Citations: | 7 - 0 self |

### BibTeX

@MISC{Urbach_efficient2-d,

author = {Erik R. Urbach and Associate Member and Michael H. F. Wilkinson and Senior Member},

title = {Efficient 2-D Grayscale Morphological Transformations With Arbitrary Flat Structuring Elements},

year = {}

}

### OpenURL

### Abstract

Abstract—An efficient algorithm is presented for the computation of grayscale morphological operations with arbitrary 2-D flat structuring elements (S.E.). The required computing time is independent of the image content and of the number of gray levels used. It always outperforms the only existing comparable method, which was proposed in the work by Van Droogenbroeck and Talbot, by a factor between 3.5 and 35.1, depending on the image type and shape of S.E. So far, filtering using multiple S.E.s is always done by performing the operator for each size and shape of the S.E. separately. With our method, filtering with multiple S.E.s can be performed by a single operator for a slightly reduced computational cost per size or shape, which makes this method more suitable for use in granulometries, dilation-erosion scale spaces, and template matching using the hit-or-miss transform. The discussion focuses on erosions and dilations, from which other transformations can be derived. Index Terms—Dilation, dilation-erosion scale spaces, erosion, fast algorithm, hit-or-miss transform, mathematical morphology, multiscale analysis. I.

### Citations

53 | Scale-space properties of the multiscale morphological dilation-erosion
- Jackway, Deriche
- 1996
(Show Context)
Citation Context ... used, and ii) application of a single operator using many different S.E.s can be computed somewhat more efficiently, which may be useful for granulometries [2], [3] and erosion-dilation scale spaces =-=[19]-=-. Compared to Van Droogenbroeck and Talbot’s method, it has the further advantage that it also works on floating point data, which is common for images originating from a Fourier transformation, such ... |

36 |
A fast algorithm for local minimum and maximum filters on rectangular and octagonal kernels
- Herk
- 1992
(Show Context)
Citation Context ...[14]. Usually, these efficient algorithms can only be used for binary images [5]–[7], [9], [12], [13], or they are limited to shapes that can (efficiently) be decomposed into a series of linear S.E.s =-=[8]-=-, [11], [14]. Efficient implementations for specialized hardware have also been studied extensively, such as the decomposition of arbitrary shapes into 3 3 blocks [10]. A recent overview of efficient ... |

32 | Recursive implementation of erosions and dilations along discrete lines at arbitrary angles
- Soille, Breen, et al.
- 1996
(Show Context)
Citation Context ... Usually, these efficient algorithms can only be used for binary images [5]–[7], [9], [12], [13], or they are limited to shapes that can (efficiently) be decomposed into a series of linear S.E.s [8], =-=[11]-=-, [14]. Efficient implementations for specialized hardware have also been studied extensively, such as the decomposition of arbitrary shapes into 3 3 blocks [10]. A recent overview of efficient algori... |

27 |
Directional morphological filtering
- Soille, Talbot
- 2001
(Show Context)
Citation Context ...ed at all. In the binary case, efficient algorithms for some 2-D shapes like circles do exist, but these cannot efficiently be extended to the grayscale case, for which polygonal approximations [11], =-=[16]-=- of circles usually are used instead. For larger circles these approximations tend be either too coarse or Manuscript received January 16, 2007; revised October 2, 2007. The associate editor coordinat... |

21 | Fast computation of morphological operations with arbitrary structuring elements
- Droogenbroeckb, Talbot
- 1996
(Show Context)
Citation Context ...mized version of the direct algorithm (like openCV or Matlab) or fast but limited to rectangular S.E.s (like Adobe Photoshop CS2). Olena [17], which uses the algorithm by Van Droogenbroeck and Talbot =-=[1]-=-, is one of the few exceptions that is faster and can handle arbitrary S.E.s. Van Droogenbroeck and Talbot [1] proposed an efficient algorithm for computing morphological operations with arbitrary 2-D... |

19 |
A contour processing method for fast binary neighbourhood operations
- Vliet, Verwer
- 1988
(Show Context)
Citation Context ...erators in mathematical morphology and have become common tools for both image filtering and analysis [3], [4] of binary and grayscale images, especially since the development of efficient algorithms =-=[5]-=-–[14]. Usually, these efficient algorithms can only be used for binary images [5]–[7], [9], [12], [13], or they are limited to shapes that can (efficiently) be decomposed into a series of linear S.E.s... |

17 |
Morphological transformations of binary images with arbitrary structuring elements
- Vincent
- 1991
(Show Context)
Citation Context ...ing and analysis [3], [4] of binary and grayscale images, especially since the development of efficient algorithms [5]–[14]. Usually, these efficient algorithms can only be used for binary images [5]–=-=[7]-=-, [9], [12], [13], or they are limited to shapes that can (efficiently) be decomposed into a series of linear S.E.s [8], [11], [14]. Efficient implementations for specialized hardware have also been s... |

16 | Efficient dilation, erosion, opening and closing algorithms
- Gil, Kimmel
- 2000
(Show Context)
Citation Context ...ors in mathematical morphology and have become common tools for both image filtering and analysis [3], [4] of binary and grayscale images, especially since the development of efficient algorithms [5]–=-=[14]-=-. Usually, these efficient algorithms can only be used for binary images [5]–[7], [9], [12], [13], or they are limited to shapes that can (efficiently) be decomposed into a series of linear S.E.s [8],... |

15 |
Decomposition of arbitrarily shaped morphological structuring elements
- Park, Chin
- 1995
(Show Context)
Citation Context ...sed into a series of linear S.E.s [8], [11], [14]. Efficient implementations for specialized hardware have also been studied extensively, such as the decomposition of arbitrary shapes into 3 3 blocks =-=[10]-=-. A recent overview of efficient algorithms for morphological operators with linear S.E. and 2-D S.E. decompositions can be found in [15]. All methods based on decomposition of 2-D S.E.s into linear S... |

12 |
Size distributions for multivariate morphological granulometries: Texture classification and statistical properties
- Batman, Dougherty
- 1997
(Show Context)
Citation Context ...erators [2] like dilation and erosion with structuring elements (S.E.) are the most fundamental operators in mathematical morphology and have become common tools for both image filtering and analysis =-=[3]-=-, [4] of binary and grayscale images, especially since the development of efficient algorithms [5]–[14]. Usually, these efficient algorithms can only be used for binary images [5]–[7], [9], [12], [13]... |

12 |
Erosion and dilation of binary images by arbitrary structuring elements using interval coding
- Liang, Piper, et al.
- 1989
(Show Context)
Citation Context ...tional efficiency of erosions is by reducing the number of redundant comparisons performed by a direct implementation of its definition A. Data Structure Much in the same way as in the binary case in =-=[6]-=-, our algorithm decomposes an arbitrary S.E. into a series of chords, i.e., runs of foreground pixels of maximum extent, as demonstrated in Fig. 1 for a letter H. Each chord can be considered as a sin... |

12 |
Morphological erosions and openings: Fast algorithms based on anchors
- Droogenbroeck, Buckley
- 2005
(Show Context)
Citation Context ... such as the decomposition of arbitrary shapes into 3 3 blocks [10]. A recent overview of efficient algorithms for morphological operators with linear S.E. and 2-D S.E. decompositions can be found in =-=[15]-=-. All methods based on decomposition of 2-D S.E.s into linear S.E.s share the same limitation: many shapes either cannot be decomposed efficiently or they cannot be decomposed at all. In the binary ca... |

6 | Generic implementation of morphological image operators
- Darbon, Géraud, et al.
- 2002
(Show Context)
Citation Context ...found to be either quite slow, being based on an processor optimized version of the direct algorithm (like openCV or Matlab) or fast but limited to rectangular S.E.s (like Adobe Photoshop CS2). Olena =-=[17]-=-, which uses the algorithm by Van Droogenbroeck and Talbot [1], is one of the few exceptions that is faster and can handle arbitrary S.E.s. Van Droogenbroeck and Talbot [1] proposed an efficient algor... |

4 | Image Analysis and Mathematical Morphology, 2 nd ed - Serra - 1982 |

4 | Decomposition of arbitrarily shaped binary morphological structuring elements using genetic algorithms
- Anelli, Broggi, et al.
- 1998
(Show Context)
Citation Context ...alysis [3], [4] of binary and grayscale images, especially since the development of efficient algorithms [5]–[14]. Usually, these efficient algorithms can only be used for binary images [5]–[7], [9], =-=[12]-=-, [13], or they are limited to shapes that can (efficiently) be decomposed into a series of linear S.E.s [8], [11], [14]. Efficient implementations for specialized hardware have also been studied exte... |

3 |
Hierarchical algorithms for morphological image processing
- Liang, Wong
- 1993
(Show Context)
Citation Context ...nd analysis [3], [4] of binary and grayscale images, especially since the development of efficient algorithms [5]–[14]. Usually, these efficient algorithms can only be used for binary images [5]–[7], =-=[9]-=-, [12], [13], or they are limited to shapes that can (efficiently) be decomposed into a series of linear S.E.s [8], [11], [14]. Efficient implementations for specialized hardware have also been studie... |

2 |
A fast implementation of 3-D binary morphological transformations
- Nikopoulos, Pitas
- 2000
(Show Context)
Citation Context ... [3], [4] of binary and grayscale images, especially since the development of efficient algorithms [5]–[14]. Usually, these efficient algorithms can only be used for binary images [5]–[7], [9], [12], =-=[13]-=-, or they are limited to shapes that can (efficiently) be decomposed into a series of linear S.E.s [8], [11], [14]. Efficient implementations for specialized hardware have also been studied extensivel... |

1 |
Mathematical morphology: Star/galaxy differentiation and galaxy morphology classification,” Publications of the Astronomical Society of Australia, preprint
- Moore, Pimbblet, et al.
(Show Context)
Citation Context ...rs [2] like dilation and erosion with structuring elements (S.E.) are the most fundamental operators in mathematical morphology and have become common tools for both image filtering and analysis [3], =-=[4]-=- of binary and grayscale images, especially since the development of efficient algorithms [5]–[14]. Usually, these efficient algorithms can only be used for binary images [5]–[7], [9], [12], [13], or ... |

1 | Efficient 2-d grayscale dilations and erosions with arbitrary flat structuring elements
- Urbach, Wilkinson
(Show Context)
Citation Context ...quest: i) our proposed (“Urbach-Wilkinson” or UW) algorithm, ii) Van Droogenbroeck and Talbot (DT) [1] for arbitrary 2-D S.E.s, and iii) Gil and Kimmel (GK) for linear S.E.s. In a preliminary version =-=[18]-=-, we presented a new method for performing morphological operators with any 2-D flat structuring element that always outperforms existing algorithms for 1057-7149/$25.00 © 2007 IEEE2 IEEE TRANSACTION... |

1 |
S’06–A’07) received the M.Sc. degree in computer science from the Institute of Mathematics and Computing Science, University of Groningen (RUG), Groningen, The Netherlands, in 2002, where he worked on the connected morphological operators for scale and sh
- pp
- 1993
(Show Context)
Citation Context ... comparisons are computed once and stored in an auxiliary array, which can be reused for filtering with all succeeding S.E.s. Note that for linear S.E.s the opening trees of Vincent are more suitable =-=[21]-=- Like the existing method of DT [1] arbitrary S.E.s can be used. However, our method always outperforms the existing method, especially when images with higher bit depth, as common in applications suc... |