izpis_h1_title_alt

Učenje konvolucijskih nevronskih mrež iz sintetičnih podatkov na primeru detekcije rok
ID Aljaž, Barbara (Author), ID Čehovin Zajc, Luka (Mentor) More about this mentor... This link opens in a new window

.pdfPDF - Presentation file, Download (3,04 MB)
MD5: DB8D4BA3DF6BCD45FE77340897FEB7E0

Abstract
Za učenje konvolucijskih nevronskih mrež je potrebna velika količina podatkov, ki jih je potrebno pridobiti in anotirati. Pogosto se za povečanje učnih zbirk uporabljajo različne augmentacije, mi pa smo v tem diplomskem delu raziskali možnost uporabe umetno generiranih podatkov. Ustvarili smo jih na podlagi tridimenzionalnega modela in parametre, ki so vplivali na zajete slike, nadzorovali avtomatsko. Delovali smo na primeru zaznave človeških rok in detektor preizkusili na dveh zbirkah realnih slik v okviru scenarija brezdotične interakcije med človekom in računalnikom. Primerjali smo ga z detektorjem, naučenim iz realističnih podatkov in analizirali razlike. Rezultati predstavljenega eksperimenta so obetavni in nakazujejo več možnosti za nadaljnji razvoj take vrste učenja.

Language:Slovenian
Keywords:računalniški vid, konvolucijske nevronske mreže, YOLO, umetno generirani podatki
Work type:Bachelor thesis/paper
Organization:FRI - Faculty of Computer and Information Science
Year:2018
PID:20.500.12556/RUL-102683 This link opens in a new window
Publication date in RUL:06.09.2018
Views:1163
Downloads:285
Metadata:XML RDF-CHPDL DC-XML DC-RDF
:
Copy citation
Share:Bookmark and Share

Secondary language

Language:English
Title:Using synthetic data to train convolutional neural networks for the case of hand detection
Abstract:
Convolutional neural networks require a large amount of data for training that need to be collected and annotated. Methods used to enlarge learning dataset usually include different augmentations, but in this thesis we researched the possibility of using artificially generated data samples. We created them using a three dimensional model and automatically controlled parameters that influenced captured images. We worked on the example of human hand detection and evaluated our detector on two datasets of real images for a touch-less interface human-computer interaction scenario. We compared it with a detector trained on real life data and analyzed the differences. Results of the experiment are promising and present many opportunities for further development of such training technique.

Keywords:computer vision, convolutional neural networks, YOLO, synthetic data

Similar documents

Similar works from RUL:
Similar works from other Slovenian collections:

Back