“You can't solve a problem with the same logic that created it”
Theses subject & Main research axes :
“Common Parallelization Strategy of Topological Operators On Shared Memory Parallel Machines.”
We deal with image segmentation and image analysis by two approaches: the first is algorithmic and the second is architectural. We study the parallelization of image processing algorithm based upon topological operators to devise a common parallelization strategy of these operators on multi-core multi-thread platform. We specifically focus our research on shared memory parallel machines. In reality, although the cost of communication (Memory-processor and inter-processors) is high enough, shared memory architectures meet our needs for different reasons: (i) These architectures have the advantage of allowing immediate sharing of data with is very helpful in the conception of any parallelization strategy (ii) They are non-dedicated architecture using standard component (processor, memory...) so economically reliable (iii) They also offer some flexibility of use in many application areas, particular image processing.
Two "Main Axes" define search operations and gather our motivation to achieve these scientific objectives: (i) 1st axe: “Study topological operators and propose parallel algorithms based upon these transformations.” (ii) 2nd axe: “Through studied algorithms, propose a common parallelization strategy suited for SMP machines.”
Start Date : 01 October 2008
End Date : 30 October 2011
Duration : 3 years
My Team : Algorithms, architectures, image analysis and computer graphics UMR 8049 LIGM
PhD TopicsView more presentations from MAHMOUDI.
PhD Supervisor :
Professeur MOHAMED AKIL - ESIEE Engineering (akilm@esiee.fr)
.
Publications Journal and conference papers :
[1] Mahmoudi, R. and Akil, M., “Concurrent Computation Of Topological Watershed On Shared Memory Parallel Machines”, [TO APPEAR - JOURNAL]
Abstract : “The watershed transform is the method of choice for image segmentation in the field of mathematical morphology. We present a new parallel topological watershed algorithm. Proposed algorithm allows a parallel computing without a prior calculation of watershed minima. To our knowledge no similar algorithm exists in the literature. We introduce an adapted parallelization strategy called split, distribute and merge (SDM) strategy which allows efficient parallelization of a large class of topological operators including, mainly, l-leveling, skeletonization and crest restoring algorithms. To achieve a good speedup, we cared about coordination of threads. Distributed work during thinning process is done by a variable number of threads. Tests on 2D grayscale image (640*640), using shared memory parallel machine (SMPM) with 8 CPU cores (2× Xeon E5405 running at frequency of 2 GHz), showed an enhancement of 6.1.”
Illustration 1 :
Slides :View more presentations from RMwebsite._____________________________________________________________
[2] Mahmoudi, R., Akil, M. and Couprie M., “Parallel 3D Skeletonization Algorithm On SMP machines”, [IN PROGRESS]
In this project we study 3D skeletonization algorithms [Bernard, Couprie 2007]. We show how, applying SDM strategy, one can parallelize and optimize this algorithm and transform it to obtain a proven parallelized and optimized algorithm. To achieve this goal, we present new method for 3D image splitting to solve banks problems in order to preserve topological features of a skeleton : homotopy, medial and thinness.
Keywords: parallel skeletonization; topology preservation; SMP machines; Multi-core / multithread architecture.
Illustration 1 :
SlidesView more presentations from RMwebsite.
[3] Mahmoudi, R., and Akil, M., “Image segmentation based upon topological operators: real-time implementation case study”, IS&T/SPIE Electronic Imaging conf., Paper 7244-1, Volume 7244, 2008, San Jose, California, USA
Abstract : “In miscellaneous applications of image treatment, thinning and crest restoring present a lot of interests. Recommended algorithms for these procedures are those able to act directly over greyscales images while preserving topology. But their strong consummation in term of time remains the major disadvantage in their choice. In this paper we present an efficient hardware implementation on RISC processor of two powerful algorithms of thinning and crest restoring developed by our team. Proposed implementation enhances execution time. A chain of segmentation applied to medical imaging will serve as a concrete example to illustrate the improvements brought thanks to the optimization techniques in both algorithm and architectural levels. The particular use of the SSE instruction set relative to the X86_32 processors (PIV 3.06 GHz) will allow a best performance for real time processing: a cadency of 33 images (512*512) per second is assure.”
Illustration 1 :
Image Segmentation ChainView more presentations from MAHMOUDI.____________________________________________________________
[4] Mahmoudi, R., Akil, M., and Matas, P., “Parallel Image Thinning through Topological Operators on Shared Memory Parallel Machines”, 43rd Asilomar Conf. on Signals, Systems, and Computers., Paper 1327,2009, Pasific Grove, California, USA
Abstract : “We present a concurrent implementation of a powerful topological thinning operator. This operator is able to act directly over grayscale image without modifying their topology. We introduce a new parallelization methodology combining SDMstrategy and thread's coordination basis which allows efficient parallelism for a large class of topological operators including skletonisation, crest restoring, 2D and 3D object smoothing and watershed. Distributed work during thinning process is done by a variant number of threads. Tests on 2D grayscale image (512*512), using shared memory parallel machines (SMPM) equipped with an octo-core processor (Xeon E5405 running at a cadency of 2Ghz), showed an enhancement of 6.2 with a maximum achieved cadency of 125images/s using 8 thread
Illustration 1 :
Slides :View more presentations from RMwebsite.[5] Mahmoudi, R. and Akil, M., “Enhanced Computation Method Of Topological Smoothing On SMP Machines”, [SUBMITTED IN EURASIP JOURNAL – Real Time Image processing on Multi-Cores FPGA-based Platforms]
Abstract : “ Smoothing is a fundamental step to reduce noise and prepare the image for subsequent processing such as segmentation. We present a new concurrent method for smoothing 2D object in binary case. Unlike existing methods, proposed method provides a parallel computation while preserving the topology. These two characteristics are ensured by the exclusive combination between homotopic transformations defined in the framework of digital topology and parallel Euclidean distance computation. We introduce an adapted parallelization strategy called split, distribute and merge (SDM) strategy which allows efficient parallelization of a large class of topological operators including, mainly, l-leveling, skeletonization, crest restoring and watershed algorithms. To achieve a good speedup, we cared about coordination of threads. Distributed work during smoothing process is done by a variable number of threads. Tests on 2D grayscale image (512*512), using shared memory parallel machine (SMPM) with 8 CPU cores (2× Xeon E5405 running at frequency of 2 GHz), showed an enhancement of 5.2.”
Illustration 1:
Illustration 2:
SlidesView more presentations from RMwebsite._____________________________________________________________
[6] Mahmoudi, R. and Akil, M., “Classification Of Parallel Thinning Algorithms On SMP Machines”, [IN PROGRESS]
Abstract : In this project we have made a selection and a classification of the best five parallel thinning algorithms that preserves topology [Bernard, Manzanera 1999][ Jang, Chin 1993][ Eckhardt, Maderlechner 1993][ Guo, Hall 1992][ Hall 1989]. In the framework of SDM strategy application, we present two different methods of image splitting that allows efficient parallel implementation of these algorithms. Evaluation tests of obtained version on SMP machines covers number of instructions, speedup, efficiency and memory consumption.
Illustration 1:
[Under constraction]
SlidesView more presentations from RMwebsite.______________________________________________________________
[7] Mahmoudi, R. and Akil, M., “Real Time Topological Image Smoothing On Shared Memory Machines”, IS&T/SPIE Electronic Imaging conf., Paper 7871-9, Volume 7247, 2011, San Francisco, California, USA
Abstract :
Illustration 1:
SlidesView more presentations from RMwebsite.
Publications : Invited Confirences :
[8] Mahmoudi, R. and Akil, M., "Lissage homotopique parallèle sur des architectures multicoeurs à mémoire partagée" AMINA Conf and workshop, Paper 60, 2010, Monastir, TN (9 p.)
Abstract :
Illustration 1:
Illustration 2:
SlidesView more presentations from RMwebsite.
