Applying parallel design techniques to template matching with GPUs

Authors:
Robert Finis Anderson;J. Steven Kirtzic;Ovidiu Daescu
Affiliations:
University of Texas at Dallas, Richardson, TX;University of Texas at Dallas, Richardson, TX;University of Texas at Dallas, Richardson, TX
Venue:
VECPAR'10 Proceedings of the 9th international conference on High performance computing for computational science
Year:
2010

Citing 13
Cited 0

Data parallel algorithms

Communications of the ACM - Special issue on parallelism
Image quilting for texture synthesis and transfer

Proceedings of the 28th annual conference on Computer graphics and interactive techniques
Introduction to Parallel Computing

Introduction to Parallel Computing
Face Recognition: Features Versus Templates

IEEE Transactions on Pattern Analysis and Machine Intelligence
Real-Time Pattern Matching Using Projection Kernels

IEEE Transactions on Pattern Analysis and Machine Intelligence
Speeding-up NCC-Based Template Matching Using Parallel Multimedia Instructions

CAMP '05 Proceedings of the Seventh International Workshop on Computer Architecture for Machine Perception
Face detection using template matching and skin-color information

Neurocomputing
Larrabee: a many-core x86 architecture for visual computing

ACM SIGGRAPH 2008 papers
Robust Real-Time Pattern Matching Using Bayesian Sequential Hypothesis Testing

IEEE Transactions on Pattern Analysis and Machine Intelligence
Full-Search-Equivalent Pattern Matching with Incremental Dissimilarity Approximations

IEEE Transactions on Pattern Analysis and Machine Intelligence
A suboptimal lossy data compression based on approximate pattern matching

IEEE Transactions on Information Theory
On Dictionary Adaptation for Recurrent Pattern Image Coding

IEEE Transactions on Image Processing
GPU accelerated 3D face registration / recognition

ICB'07 Proceedings of the 2007 international conference on Advances in Biometrics

Quantified Score

Hi-index	0.00

Visualization

Abstract

Designing algorithms for data parallelism can create significant gains in performance on SIMD architectures. The performance of General Purpose GPUs can also benefit from careful analysis of memory usage and data flow due to their large throughput and system memory bottlenecks. In this paper we present an algorithm for template matching that is designed from the beginning for the GPU architecture and achieves greater than an order of magnitude speedup over traditional algorithms designed for the CPU and reimplemented on the GPU. This shows that it is not only desirable to adapt existing algorithms to run on GPUs, but also that future algorithms should be designed with the GPU architecture in mind.