A. Rosenfeld, J. Ornelas, and Y. Hung, "Hough transform algorithms for mesh-connected SIMD parallel processors", Computer Vision, Graphics, and Image Processing, 41, 1988, pp 293-305.  Home/Search   Document Not in Database   Summary   Related Articles   Check

....threshold value define likely edges. The pair (r, j) defines the unique line L whose normal has length r and angle q j . On a single processor computer, the Hough transform is easily computed in O (N p) time. Parallel algorithms for mesh connected computers have been proposed by Rosenfeld et al. [ROSE88], Cypher et al. CYPH87] Guerra and Hambrush [GUER87] and Silberberg [SILB85] The algorithm of Cypher et al. uses a pipelined technique and has complexity O (N p) on an NN mesh. Fisher and Highnam [FISH87] consider a scan line array. Their algorithm has time complexity O (N p) and is ....

A. Rosenfeld, J. Ornelas, and Y. Hung, "Hough transform algorithms for mesh-connected SIMD parallel processors", Computer Vision, Graphics, and Image Processing, 41, 1988, pp 293-305.

....to the p angles Oj p (j 1) Since Oj is in the range (0, r] and since 0 x, y N, i is in the range [ x N, x N] Parallel algorithms to compute the Hough transform have been developed for several architec tures. Chaung and Li [2] and Li et a . 23] do this for systolic arrays; Rosenreid et a . [34], Kannan and Chaung [17] Cypher et a . 4] Guerra and Hambrusch [8] and Silberberg [37] consider mesh computers; Fisher and Highham [6] use a scan line array; Ibrahim et a . 10] uses a SIMD tree; Li et a . 21, 22] Maresca et a . 25] use a polymorphic torus; Ranka and Sahni [32] use hypercube ....

A. Rosenreid, J. Ornelas, Jr., and Y. Hung. Hough transform algorithms for mesh-connected SIMD parallel processors. Computer Vision, Graphics, and Image Processing, 41:293-305, 1988.

.... VLSI implementations have been proposed in the literature [12] 17] 19] In particular, a number of parallel Hough transform algorithms have been designed for different parallel architectures, including tree [7] systolic array [1] 16] line array [4] pyramid [9] mesh [2] 5] 10] 22] [27], 30] reconfigurable mesh [8] and hypercube [23] Recently, several constant time algorithms for computing the Hough transform have been proposed for the reconfigurable mesh model [11] 24] In the above algorithms, the data buses in the reconfigurable mesh model are O(log N) bit wide. More ....

A. Rosenfeld, J. Ornelas, and Y. Hung, "Hough transform algorithms for mesh-connected SIMD parallel processors," Comput. Vis., Graph. Image Process., vol. 41, pp. 293--305, 1988.

....to be taken in a fast way. In these cases, the use of parallel computers is necessary. The parallelization of the classical HT can be carried out using loop projection techniques. These solutions have been used both in multiprocessors with a coarse grain [6] and in array processors with fine grain [7, 8, 9, 10]. The main problem in these implementations is the use of a common polling parameter space that causes contention problems in shared memory parallel computers and a large number of messages in distributed memory parallel computers. The parallelization of algorithms based on a focusing process, i. ....

A. Rosenfeld, J. Ornelas and Y. Hung. "Hough transform algorithms for meshconnected SIMD parallel processor". J. Computer Vision Graphics Imagen Processing, Vol. 41, pp. 293-305, 1988.

....study of the parallelization of the Hough transform has been performed previously. For the detection of lines, Ranka and Sahni designed efficient algorithms to compute this specialized version of the Hough transform on MIMD and SIMD hypercube multicomputers [6] Other research by Rosenfeld et al. [7], Cypher and Sanz [4] and Guerra and Hambrusch [5] discusses mesh connected multicomputers as a means for calculating the Hough transform. Work funded in part by NSF Grant CDA 9200577 for the 1992 Research Experiences for Undergraduates (REU) Site Program at the Northeast Parallel ....

A. Rosenfeld, J. Ornelas and Y. Hung, "Hough Transform Algorithms for Mesh-connected SIMD Parallel Processors," CVGIP 41, No. 3, pp. 293--305, 1988.

....large it is essential to be able to speed up the computation of the HT. Parallel processing was proposed as a possible solution. Parallel algorithms for line detection using the HT have been designed for different parallel architectures including tree [13] linear array [6] pyramid [15] mesh [4, 7, 17, 28, 32, 33] and hypercube [27] A number of algorithms for line detection using the HT of an N Theta N image on reconfigurable networks have also been proposed in the literature. The algorithms given in [16, 23] have O(p log(N=p) time complexity and use an N ThetaN reconfigurable mesh. In [16] the time ....

Rosenfeld, A., Ornelas, J., Hung, Y., Hough transform algorithms for mesh-connected SIMD parallel processors, Computer Vision, Graphics and Image Processing, 41, 1988, 293-305.

....mapping of Hough Transform algorithm described in Chapter 5 where a ae value is computed for a given with Equation 5.1 and the corresponding cell is incremented, the global shifting and summation of data from all part of the mesh are required. It has been shown by Silberberg [29] and Rosenfeld [30] that such a mapping consequently results in a poor speed up. Therefore, an alternative approach has to be taken to achieve efficient parallel implementation of the Hough Transform. As a different interpretation of the Hough Transform, it can be seen that the Hough Transform accumulates the number ....

A. Rosenfeld, J. Ornelas, and Hung Yubin, Hough transform algorithms for mesh-connected SIMD parallel processors, Computer Vision Graphics and Image Processing, 41, 1988.

.... The most important problem presented by these fast algorithms is their low regularity and parallelism [22] whereas the traditional algorithm has a strong implicit parallelism which can be conveniently used in shared memory [11] or distributed memory multiprocessors (linear array [20] mesh [44], 5] 19] 3] 13] hypercube [7] 45] binary tree [9] The algorithm associated with the Hough transform presents a high level of regularity and locality. These properties make it appropriate for the VLSI design of a specific purpose architecture which permits high speed computation. ....

A. Rosenfeld, J. Ornelas and Y. Hung. "Hough transform algorithms for mesh-connected SIMD parallel processor". J. Computer Vision Graphics Imagen Processing, Vol. 41, pp. 293-305, 1988.

....with each processor performing the whole algorithm. Due to its inherent (sub image) nature the R sHT is best suited for a Data Parallelism implementation. The Hough Transform and some of its variations have been theoretically and practically implemented on a number of parallel systems, e.g. [7], and [8] Here, we consider the practical implementation of the R sHT on transputer arrays and farms. Data routing amongst network processors is a major efficiency factor. For transputerbased real time applications, the sequences involving the distribution and collection of raw data from the ....

Rosenfeld A., Ornelas J., and Hung Y. Hough transform algorithms for mesh-connected simd parallel processors. Computer Vision, Graphics and Image Processing, 41:293--305, 1988.

....and hence good candidates for benefiting from parallelism. Most work has been on SIMD implementations of low level tasks (Kittler and Duff, 1985) where ease of problem decomposition enables good performance from simple algorithms. Intermediate vision algorithms have been implemented on both SIMD (Rosenfeld et al. 1988) and MIMD (Austin et al. 1991) architectures. As with low level activities, pixel based data eases partitioning. Parallel high level vision is relatively recent (Bhanu and Nuttall, 1989; Amini et al. 1989) Here, the main requirement is to constrain the search space which grows with both scene ....

A. Rosenfeld, J. Ornelas, and Y. Hung. Hough transform algorithms for mesh-connected SIMD parallel processors. Computer Vision, Graphics and Image Processing, 41:293-- 305, 1988.

Online articles have much greater impact   More about CiteSeer.IST   Add search form to your site   Submit documents   Feedback  

CiteSeer.IST - Copyright Penn State and NEC