Naslov
GPU Implementation of a Medical Imaging Data Compression Algorithm

Autori
Šantl, Dino ; Đurasević, Marko ; Knezović, Josip

Vrsta, podvrsta i kategorija rada
Radovi u zbornicima skupova, cjeloviti rad (in extenso), znanstveni

Izvornik
IT Systems 2013, e-Health 2013 Electronic Proceedings / Žagar, Martin ; Knezović, Josip ; Mlinarić Hrvoje ; Hofman Daniel ; Kovač, Mario - Zagreb : Fakultet elektrotehnike i računarstva Sveučilišta u Zagrebu, 2013, 317-321

ISBN
978-953-184-189-4

Skup
IT Systems 2013 and e-Health 2013

Mjesto i datum
Bol, Hrvatska, 12.09.2013. - 14.09.2013

Vrsta sudjelovanja
Predavanje

Vrsta recenzije
Međunarodna recenzija

Ključne riječi
Medical Visual Data Compression; General Purpose GPU Programming; Lossless Coding Parallelization; Data-Intensive Computing

Sažetak
In this paper an algorithm for compressing medical images on GPUs is described and its energy consumption and computational performance analyzed. The objectives of this implementation are to reduce the demands of storing medical pictures, improve the execution time and reduce the consumption of electrical energy. The compression algorithm is based on a computationally complex prediction model of pixels and a contextual coding of the given prediction errors. The above implementation is compared with a serial implementation and with an implementation in a stream programming model in terms of execution time and electrical energy consumption. Experimental results show improvements in execution speed and savings in consumption of electrical energy using the proposed implementation

Izvorni jezik
Engleski

Znanstvena područja
Računarstvo

POVEZANOST RADA