Abstract
Non-Local Means (NLM) algorithm is widely considered as a state-of-the-art denoising filter in many research fields. High computational complexity led to implementations on Graphic Processor Unit (GPU) architectures, which achieve reasonable running times by filtering, slice-by-slice, 3D datasets with a 2D NLM approach. Here we present a fully 3D NLM implementation on a multi-GPU architecture and suggest its high scalability. The performance results we discuss encourage the coding of further filter improvements and the investigation of a large spectrum of applicative scenarios.
| Original language | English |
|---|---|
| Title of host publication | 2013 Federated Conference on Computer Science and Information Systems, FedCSIS 2013 |
| Pages | 495-498 |
| Number of pages | 4 |
| Publication status | Published - 2013 |
| Event | 2013 Federated Conference on Computer Science and Information Systems, FedCSIS 2013 - Krakow, Poland Duration: Sep 8 2013 → Sep 11 2013 |
Other
| Other | 2013 Federated Conference on Computer Science and Information Systems, FedCSIS 2013 |
|---|---|
| Country | Poland |
| City | Krakow |
| Period | 9/8/13 → 9/11/13 |
Fingerprint
ASJC Scopus subject areas
- Computer Science (miscellaneous)
- Information Systems
Cite this
3D Non-Local Means denoising via multi-GPU. / Palma, Giuseppe; Comerci, Marco; Alfano, Bruno; Cuomo, Salvatore; De Michele, Pasquale; Piccialli, Francesco; Borrelli, Pasquale.
2013 Federated Conference on Computer Science and Information Systems, FedCSIS 2013. 2013. p. 495-498 6644045.Research output: Chapter in Book/Report/Conference proceeding › Conference contribution
}
TY - GEN
T1 - 3D Non-Local Means denoising via multi-GPU
AU - Palma, Giuseppe
AU - Comerci, Marco
AU - Alfano, Bruno
AU - Cuomo, Salvatore
AU - De Michele, Pasquale
AU - Piccialli, Francesco
AU - Borrelli, Pasquale
PY - 2013
Y1 - 2013
N2 - Non-Local Means (NLM) algorithm is widely considered as a state-of-the-art denoising filter in many research fields. High computational complexity led to implementations on Graphic Processor Unit (GPU) architectures, which achieve reasonable running times by filtering, slice-by-slice, 3D datasets with a 2D NLM approach. Here we present a fully 3D NLM implementation on a multi-GPU architecture and suggest its high scalability. The performance results we discuss encourage the coding of further filter improvements and the investigation of a large spectrum of applicative scenarios.
AB - Non-Local Means (NLM) algorithm is widely considered as a state-of-the-art denoising filter in many research fields. High computational complexity led to implementations on Graphic Processor Unit (GPU) architectures, which achieve reasonable running times by filtering, slice-by-slice, 3D datasets with a 2D NLM approach. Here we present a fully 3D NLM implementation on a multi-GPU architecture and suggest its high scalability. The performance results we discuss encourage the coding of further filter improvements and the investigation of a large spectrum of applicative scenarios.
UR - http://www.scopus.com/inward/record.url?scp=84892514755&partnerID=8YFLogxK
UR - http://www.scopus.com/inward/citedby.url?scp=84892514755&partnerID=8YFLogxK
M3 - Conference contribution
AN - SCOPUS:84892514755
SN - 9781467344715
SP - 495
EP - 498
BT - 2013 Federated Conference on Computer Science and Information Systems, FedCSIS 2013
ER -