An Efficient Hardware Architecture without Line Memories for Morphological Image Processing

Clienti, Christophe; Bilodeau, Michel; Beucher, Serge

doi:10.1007/978-3-540-88458-3_14

Cited by 16 publications

(13 citation statements)

References 5 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…Where decompositions have been used, they have generally been restricted to the series decomposition (see for example [4]). However, the size and shape of the SEs that can be created by series decomposition is limited.…”

Section: B Prior Workmentioning

confidence: 99%

Efficient implementation of greyscale morphological filters

Bailey

2010

2010 International Conference on Field-Programmable Technology

View full text Add to dashboard Cite

Morphological filters are often implemented using a series decomposition. This paper presents a parallel decomposition that is able to exploit separability. By maximising the reuse of hardware between the parallel filters, a novel computationally efficient filter structure may be derived. Results show that such filters may be implemented on a Virtex-5 FPGA with pixel clock rates approaching 1 GHz.

show abstract

Section: B Prior Workmentioning

confidence: 99%

Efficient implementation of greyscale morphological filters

Bailey

2010

2010 International Conference on Field-Programmable Technology

View full text Add to dashboard Cite

show abstract

“…In [26], Soille et al extended this work to arbitrary-oriented openings. In [27], Clienti et al improved the HGW algorithm by removing the image backward scanning to reduce latency.…”

Section: Opening Algorithmsmentioning

confidence: 99%

“…Their implementation was tested on an Intel CPU with the SIMD-SSE2 instruction set. Clienti [27] improved the HGW algorithm and implemented 2 it on a SIMD architecture as well. Domanski et al [33] used CUDA to implement the HGW algorithm on GPU, achieving 13-33 × speedup.…”

Section: Parallel Implementationsmentioning

confidence: 99%

GPU implementation of linear morphological openings with arbitrary angle

Karas

Morard

Grandpierre

et al. 2012

J Real-Time Image Proc

View full text Add to dashboard Cite

Linear morphological openings and closings are important non-linear operators from mathematical morphology. In practical applications, many different orientations of digital line segments must typically be considered. In this paper, we (1) review efficient sequential as well as parallel algorithms for the computation of linear openings and closings, (2) compare the performance of CPU implementations of four state-of-the-art algorithms, (3) describe GPU implementations of two recent efficient algorithms allowing arbitrary orientation of the line segments, (4) propose, as the main contribution, an efficient and optimized GPU implementation of linear openings, and (5) compare the performance of all implementations on real images from various applications. From our experimental results, it turned out that the proposed GPU implementation is suitable for applications with large, industrial images, running under severe timing constraints.

show abstract

“…This algorithm, called hereafter HGW, is independent of the size of the structuring element for the computation of one-dimensional (1-D) erosions and dilations. Later, Clienti et al improved HGW algorithm by removing the backward scanning to ensure a low latency [18]. Then, algorithms have also been proposed to compute openings in only one pass of the entire image, without computing successively the erosion and the dilation.…”

Section: Introductionmentioning

confidence: 99%

One-Dimensional Openings, Granulometries and Component Trees in 𝒪(1) in Per Pixel

Morard

Dokládal

Decencière

2012

IEEE J. Sel. Top. Signal Process.

View full text Add to dashboard Cite

We introduce a new, efficient and adaptable algorithm to compute openings, granulometries and the component tree for onedimensional (1-D) signals. The algorithm requires only one scan of the signal, runs in place in O(1) per pixel, and supports any scalar data precision (integer or floating-point data). The algorithm is applied to two-dimensional images along straight lines, in arbitrary orientations. Oriented size distributions can thus be efficiently computed, and textures characterised. Extensive benchmarks are reported. They show that the proposed algorithm allows computing 1-D openings faster than existing algorithms for data precisions higher than 8 bits, and remains competitive with respect to the algorithm proposed by Van Droogenbroeck when dealing with 8-bit images. When computing granulometries, the new algorithm runs faster than any other method of the state of the art. Moreover, it allows efficient computation of 1-D component trees.

show abstract

An Efficient Hardware Architecture without Line Memories for Morphological Image Processing

Cited by 16 publications

References 5 publications

Efficient implementation of greyscale morphological filters

Efficient implementation of greyscale morphological filters

GPU implementation of linear morphological openings with arbitrary angle

One-Dimensional Openings, Granulometries and Component Trees in 𝒪(1) in Per Pixel

Contact Info

Product

Resources

About