Presentation 2009-03-06
Memory Efficient Pixel-parallel Hough Transform Circuit Using Collision-free Voting Memory Access
Masayuki GOTO, Kazuhiro NAKAMURA, Kazuyoshi TAKAGI, Naofumi TAKAGI,
PDF Download Page PDF download Page Link
Abstract(in Japanese) (See Japanese page)
Abstract(in English) The Hough transform in computer vision algorithms is a method which detects lines from an image and is used in embedded systems such as cellular phones and vehicles because of the robustness. The more the number of pixel, the more difficult it is to process Hough transform in real-time. Thus, researchs on hardware acceleration which parallelize voting operations on the angular parameter θ have been done. In this report, we propose a method for eliminating vote collisions. Collision-free voting memory access is accomplished by selecting the direction of parallelism and changing accuracy of rounding. The voting memory accesses are collision-free. A pixel-parallel Hough transform circuit using collision-free and localized voting memory access is more memory efficient than a θ-parallel circuit and a pixel-parallel circuit, and collision-free. We have evaluated the circuit by comparing clock cycles and amount of memory in the circuit of the circuit and conventional circuits. We compared 16-parallel circuits for VGA image size, and the result shows that clock cycles are almost the same and the circuit has approximately 99% less amount of memory than a θ-parallel circuit.
Keyword(in Japanese) (See Japanese page)
Keyword(in English) Hough transform / ASIC / image processing / collision-free voting / parallel Hough transform circuit
Paper # CPSY2008-101,DC2008-92
Date of Issue

Conference Information
Committee DC
Conference Date 2009/2/26(1days)
Place (in Japanese) (See Japanese page)
Place (in English)
Topics (in Japanese) (See Japanese page)
Topics (in English)
Chair
Vice Chair
Secretary
Assistant

Paper Information
Registration To Dependable Computing (DC)
Language JPN
Title (in Japanese) (See Japanese page)
Sub Title (in Japanese) (See Japanese page)
Title (in English) Memory Efficient Pixel-parallel Hough Transform Circuit Using Collision-free Voting Memory Access
Sub Title (in English)
Keyword(1) Hough transform
Keyword(2) ASIC
Keyword(3) image processing
Keyword(4) collision-free voting
Keyword(5) parallel Hough transform circuit
1st Author's Name Masayuki GOTO
1st Author's Affiliation Graduate School of Information Science, Nagoya University()
2nd Author's Name Kazuhiro NAKAMURA
2nd Author's Affiliation Graduate School of Information Science, Nagoya University
3rd Author's Name Kazuyoshi TAKAGI
3rd Author's Affiliation Graduate School of Information Science, Nagoya University
4th Author's Name Naofumi TAKAGI
4th Author's Affiliation Graduate School of Information Science, Nagoya University
Date 2009-03-06
Paper # CPSY2008-101,DC2008-92
Volume (vol) vol.108
Number (no) 464
Page pp.pp.-
#Pages 6
Date of Issue