Title:	Comparison of Performance of Optimized HSI CNN models on Desktop and Embedded Platforms
Authors:	Čech, Jiří Rozkovec, Martin
Citation:	2021 International Conference on Applied Electronics: Pilsen, 7th – 8th September 2021, Czech Republic, p. 29-32.
Issue Date:	2021
Publisher:	University of West Bohemia
Document type:	conferenceObject konferenční příspěvek
URI:	http://hdl.handle.net/11025/45560
ISBN:	978–80–261–0972–3 (Print) 978–80–261–0973–0 (Online)
ISSN:	1803–7232 (Print) 1805–9597 (Online)
Keywords:	hyperspektrální zobrazování;neuronové sítě;Xilinx FPGA;Vitis AI;DPU
Keywords in different language:	hyperspectral imaging;neural networks;Xilinx FPGA;Vitis AI;DPU
Abstract in different language:	We compare different platforms for inference of convolutional neural networks in this paper. We trained various neural networks to determine the material in the source hyperspectral cube. Then we convert them to inference format and compare the inference results. We used tools under Xilinx Vitis AI for FPGA implementation. We try to minimize the size of the proposed networks by pruning them and provide further comparisons. FPGA platforms show to be energy efficient but still slower than a graphics card in terms of performance.
Rights:	© University of West Bohemia, 2021
Appears in Collections:	Applied Electronics 2021 Applied Electronics 2021

Items in DSpace are protected by copyright, with all rights reserved, unless otherwise indicated.