Please use this identifier to cite or link to this item: https://scholarbank.nus.edu.sg/handle/10635/107809
Title: ACCELERATING REAL-TIME COMPUTER VISION ALGORITHMS ON PARALLEL HARDWARE ARCHITECTURES.
Authors: ANG ZHI PING
Keywords: FPGA, sparse recovery, real-time, embedded, GPGPU, parallel computing
Issue Date: 20-May-2014
Citation: ANG ZHI PING (2014-05-20). ACCELERATING REAL-TIME COMPUTER VISION ALGORITHMS ON PARALLEL HARDWARE ARCHITECTURES.. ScholarBank@NUS Repository.
Abstract: BASIS PURSUIT DENOISING (BPDN) IS AN OPTIMIZATION METHOD USED IN CUTTING EDGE COMPUTER VISION RESEARCH. HOSTING AN EMBEDDED BPDN SOLVER IS DESIRABLE BECAUSE ANALYSIS CAN BE PERFORMED IN REAL-TIME, BUT EXISTING SOLVERS ARE GENERALLY UNSUITABLE FOR EMBEDDED IMPLEMENTATION DUE TO EITHER POOR RUN-TIME PERFORMANCE OR LACK OF PARALLEL STRUCTURE. TO ADDRESS THE AFOREMENTIONED ISSUES, THIS THESIS PROPOSES AN EMBEDDED-FRIENDLY SOLVER WHICH DEMONSTRATES SUPERIOR RUN-TIME PERFORMANCE, HIGH RECOVERY ACCURACY AND COMPETITIVE MEMORY USAGE COMPARED TO EXISTING SOLVERS. FOR A PROBLEM WITH 5000 VARIABLES AND 500 CONSTRAINTS, THE SOLVER OCCUPIES A SMALL MEMORY FOOTPRINT OF 29 KB AND TAKES 0.14 SECONDS TO COMPLETE ON THE XILINX ZYNQ Z-7020 FPGA. THE SOLVER IS ALSO IMPLEMENTED ON THE NVIDIA TESLA M2050 GPU. THE THESIS FURNISHES AN OPTIMIZED ALGORITHM TO SOLVE THE GENERALIZED MATRIX-VECTOR MULTIPLICATION ON THE GPU, WHICH IS AT LEAST TWICE AS FAST AS CUBLAS AND 702 TIMES AS FAST AS THE FPGA IMPLEMENTATION.
URI: http://scholarbank.nus.edu.sg/handle/10635/107809
Appears in Collections:Master's Theses (Open)

Show full item record
Files in This Item:
File Description SizeFormatAccess SettingsVersion 
thesis.pdf2.64 MBAdobe PDF

OPEN

NoneView/Download

Page view(s)

219
checked on Oct 14, 2021

Download(s)

60
checked on Oct 14, 2021

Google ScholarTM

Check


Items in DSpace are protected by copyright, with all rights reserved, unless otherwise indicated.