Parallel implementations of motion estimation algorithms using OpenCL

Antti Heikkinen; Lance Fono

doi:10.1109/ICDSP.2013.6622694

Parallel implementations of motion estimation algorithms using OpenCL

Research output: Chapter in Book/Report/Conference proceeding › Conference article in proceedings › Scientific › peer-review

3 Citations (Scopus)

Abstract

Parallel processors such as Graphics processing units (GPUs) have emerged as co-processing units for central processing units (CPUs) to accelerate different applications. Open Computing Language (OpenCL) is a framework for multiprocessing in heterogeneous platforms. In this paper we focus on motion estimation which is the most time consuming task in video coding. We study two motion estimation algorithms in terms of parallel execution. We implemented the full search algorithm and the hierarchical search algorithm with OpenCL and with C code. Our measurements show that the OpenCL-based implementations of the algorithms on the GPU can achieve nearly 10 times speedup compared to the corresponding C implementation on a single CPU.

Original language	English
Title of host publication	18th International Conference on Digital Signal Processing, DSP 2013
Publisher	IEEE Institute of Electrical and Electronic Engineers
ISBN (Print)	978-1-4673-5805-7
DOIs	https://doi.org/10.1109/ICDSP.2013.6622694
Publication status	Published - 2013
MoE publication type	A4 Article in a conference publication
Event	18th International Conference on Digital Signal Processing, DSP 2013 - Santorini, Greece Duration: 1 Jul 2013 → 3 Jul 2013

Conference

Conference	18th International Conference on Digital Signal Processing, DSP 2013
Abbreviated title	DSP 2013
Country/Territory	Greece
City	Santorini
Period	1/07/13 → 3/07/13

Keywords

Hieararchical motion estimation
full search motion estimation
OpenCL
parallel
GPU
video coding

Access to Document

10.1109/ICDSP.2013.6622694

Cite this

@inproceedings{70b107407d2444e3bd129fa23cbe9a55,

title = "Parallel implementations of motion estimation algorithms using OpenCL",

abstract = "Parallel processors such as Graphics processing units (GPUs) have emerged as co-processing units for central processing units (CPUs) to accelerate different applications. Open Computing Language (OpenCL) is a framework for multiprocessing in heterogeneous platforms. In this paper we focus on motion estimation which is the most time consuming task in video coding. We study two motion estimation algorithms in terms of parallel execution. We implemented the full search algorithm and the hierarchical search algorithm with OpenCL and with C code. Our measurements show that the OpenCL-based implementations of the algorithms on the GPU can achieve nearly 10 times speedup compared to the corresponding C implementation on a single CPU.",

keywords = "Hieararchical motion estimation, full search motion estimation, OpenCL, parallel, GPU, video coding",

author = "Antti Heikkinen and Lance Fono",

note = "Project code: 75725; 18th International Conference on Digital Signal Processing, DSP 2013, DSP 2013 ; Conference date: 01-07-2013 Through 03-07-2013",

year = "2013",

doi = "10.1109/ICDSP.2013.6622694",

language = "English",

isbn = "978-1-4673-5805-7",

booktitle = "18th International Conference on Digital Signal Processing, DSP 2013",

publisher = "IEEE Institute of Electrical and Electronic Engineers",

address = "United States",

}

Heikkinen, A & Fono, L 2013, Parallel implementations of motion estimation algorithms using OpenCL. in 18th International Conference on Digital Signal Processing, DSP 2013., 6622694, IEEE Institute of Electrical and Electronic Engineers, 18th International Conference on Digital Signal Processing, DSP 2013, Santorini, Greece, 1/07/13. https://doi.org/10.1109/ICDSP.2013.6622694

Parallel implementations of motion estimation algorithms using OpenCL. / Heikkinen, Antti; Fono, Lance.

18th International Conference on Digital Signal Processing, DSP 2013. IEEE Institute of Electrical and Electronic Engineers, 2013. 6622694.

Research output: Chapter in Book/Report/Conference proceeding › Conference article in proceedings › Scientific › peer-review

TY - GEN

T1 - Parallel implementations of motion estimation algorithms using OpenCL

AU - Heikkinen, Antti

AU - Fono, Lance

N1 - Project code: 75725

PY - 2013

Y1 - 2013

N2 - Parallel processors such as Graphics processing units (GPUs) have emerged as co-processing units for central processing units (CPUs) to accelerate different applications. Open Computing Language (OpenCL) is a framework for multiprocessing in heterogeneous platforms. In this paper we focus on motion estimation which is the most time consuming task in video coding. We study two motion estimation algorithms in terms of parallel execution. We implemented the full search algorithm and the hierarchical search algorithm with OpenCL and with C code. Our measurements show that the OpenCL-based implementations of the algorithms on the GPU can achieve nearly 10 times speedup compared to the corresponding C implementation on a single CPU.

AB - Parallel processors such as Graphics processing units (GPUs) have emerged as co-processing units for central processing units (CPUs) to accelerate different applications. Open Computing Language (OpenCL) is a framework for multiprocessing in heterogeneous platforms. In this paper we focus on motion estimation which is the most time consuming task in video coding. We study two motion estimation algorithms in terms of parallel execution. We implemented the full search algorithm and the hierarchical search algorithm with OpenCL and with C code. Our measurements show that the OpenCL-based implementations of the algorithms on the GPU can achieve nearly 10 times speedup compared to the corresponding C implementation on a single CPU.

KW - Hieararchical motion estimation

KW - full search motion estimation

KW - OpenCL

KW - parallel

KW - GPU

KW - video coding

U2 - 10.1109/ICDSP.2013.6622694

DO - 10.1109/ICDSP.2013.6622694

M3 - Conference article in proceedings

SN - 978-1-4673-5805-7

BT - 18th International Conference on Digital Signal Processing, DSP 2013

PB - IEEE Institute of Electrical and Electronic Engineers

T2 - 18th International Conference on Digital Signal Processing, DSP 2013

Y2 - 1 July 2013 through 3 July 2013

ER -

Parallel implementations of motion estimation algorithms using OpenCL

Abstract

Conference

Keywords

Access to Document

Fingerprint

Cite this