GPGPU implementation of growing neural gas: application to 3D scene reconstruction
Please use this identifier to cite or link to this item:
http://hdl.handle.net/10045/34195
Title: | GPGPU implementation of growing neural gas: application to 3D scene reconstruction |
---|---|
Authors: | Orts-Escolano, Sergio | Garcia-Rodriguez, Jose | Viejo Hernando, Diego | Cazorla, Miguel | Morell, Vicente |
Research Group/s: | Robótica y Visión Tridimensional (RoViT) | Informática Industrial y Redes de Computadores |
Center, Department or Service: | Universidad de Alicante. Instituto Universitario de Investigación Informática | Universidad de Alicante. Departamento de Ciencia de la Computación e Inteligencia Artificial | Universidad de Alicante. Departamento de Tecnología Informática y Computación |
Keywords: | Growing neural gas | Parallel computing | GPU | CUDA | Multicore | 3D reconstruction | Egomotion |
Knowledge Area: | Ciencia de la Computación e Inteligencia Artificial | Tecnologías de la información y la Computación |
Issue Date: | Oct-2012 |
Publisher: | Elsevier |
Citation: | Journal of Parallel and Distributed Computing. 2012, 72(10): 1361-1372. doi:10.1016/j.jpdc.2012.05.008 |
Abstract: | Self-organising neural models have the ability to provide a good representation of the input space. In particular the Growing Neural Gas (GNG) is a suitable model because of its flexibility, rapid adaptation and excellent quality of representation. However, this type of learning is time-consuming, especially for high-dimensional input data. Since real applications often work under time constraints, it is necessary to adapt the learning process in order to complete it in a predefined time. This paper proposes a Graphics Processing Unit (GPU) parallel implementation of the GNG with Compute Unified Device Architecture (CUDA). In contrast to existing algorithms, the proposed GPU implementation allows the acceleration of the learning process keeping a good quality of representation. Comparative experiments using iterative, parallel and hybrid implementations are carried out to demonstrate the effectiveness of CUDA implementation. The results show that GNG learning with the proposed implementation achieves a speed-up of 6× compared with the single-threaded CPU implementation. GPU implementation has also been applied to a real application with time constraints: acceleration of 3D scene reconstruction for egomotion, in order to validate the proposal. |
Sponsor: | This work has been supported by grant DPI2009-07144 from Ministerio de Ciencia e Innovacion of the Spanish Government, by the University of Alicante projects GRE09-16 and GRE10-35, and Valencian Government project GV/2011/034. Experiments were made possible with a generous donation of hardware from NVDIA. |
URI: | http://hdl.handle.net/10045/34195 |
ISSN: | 0743-7315 (Print) | 1096-0848 (Online) |
DOI: | 10.1016/j.jpdc.2012.05.008 |
Language: | eng |
Type: | info:eu-repo/semantics/article |
Peer Review: | si |
Publisher version: | http://dx.doi.org/10.1016/j.jpdc.2012.05.008 |
Appears in Collections: | INV - RoViT - Artículos de Revistas INV - I2RC - Artículos de Revistas INV - AIA - Artículos de Revistas |
Files in This Item:
File | Description | Size | Format | |
---|---|---|---|---|
jdpc.pdf | Preprint (acceso abierto) | 1,7 MB | Adobe PDF | Open Preview |
jdpc-final.pdf | Versión final (acceso restringido) | 3,11 MB | Adobe PDF | Open Request a copy |
Items in RUA are protected by copyright, with all rights reserved, unless otherwise indicated.