20.500.12556/DKUM-67653 Uporaba globokega učenja s knjižnico Deeplearning4j na primeru prepoznave obrazov The use of deep learning with Deeplearning4j on the case of facial recognition V magistrskem delu smo se dotaknili področja globokega učenja, spoznali smo pristope in arhitekture algoritmov globokega učenja ter jih kategorizirali v tri skupine. V nadaljevanju smo podrobneje analizirali knjižnico Deeplearning4j, predstavili osnovne funkcionalnosti ter raziskali njene možnosti za uporabo na področju globokega učenja. V praktičnem delu smo uporabo globokega učenja s knjižnico Deeplearning4j aplicirali na primeru prepoznave obrazov. Implementirali smo dve različici konvolucijskih nevronskih mrež ter dva načina učenja – lokalno ter porazdeljeno učenje. In the master’s thesis, we briefly introduced the area of deep learning. We explored and described the approaches and architectures of deep learning algorithms and categorize them into three groups. In the following, we analyzed Deeplearning4j library, presented main features and studied possibilities of its use in the field of deep learning. In the empirical part, the use of deep learning with Deeplearning4j was applied on the case of facial recognition. We implemented two different versions of convolutional neural networks and two types of learning – local and distributed learning. strojno učenje globoko učenje Deeplearning4j prepoznava obraza machine learning deep learning Deeplearning4j facial recognition true true false G. Vrbančič Slovenski jezik Angleški jezik Magistrsko delo/naloga 2017-08-25 18:19:21 2017-10-17 15:11:28 2022-08-01 14:17:18 2019-07-11 16:26:54 2017 0 [Maribor 0 0000-00-00 NiDoloceno NiDoloceno NiDoloceno 0000-00-00 0000-00-00 0000-00-00 004.932.72'1(043.2) 20969494 URN:SI:UM:DK:JEAMJ5I1 MAG_Vrbancic_Grega_2017.pdf MAG_Vrbancic_Grega_2017.pdf 1 B52311B91BD86D0FAC371C8C5AD25E13 686489d54628cd535147db7b3d21b047f1d3f4ce775cb1b8b1579c0e746a2a8e 91360276-7c0e-11eb-bb7a-00155d0001ca 20.500.12556/dkum/959548db-e957-4b4a-896c-67c529139281 https://dk.um.si/Dokument.php?lang=slv&id=116091 Fakulteta za elektrotehniko, računalništvo in informatiko 0 0 0