Prepoznava prstov s pomočjo globokega učenja

magistrsko delo

Robert Kopušar (Author), Dušan Gleich (Mentor)

Abstract

Magistrsko delo obravnava problematiko prepoznavanja prstov na roki, s pomočjo katere lahko v ozadju upravljamo najrazličnejše naloge in procese. Delo je zasnovano kot predstavitev reševanja iste problematike s pomočjo dveh različnih pristopov in predstavitev njunih prednosti in slabosti. Z uporabo tehnologije iskanja vzorca v sliki smo se problematike lotili na direkten način in v sliki sami iskali značilnost, s pomočjo katere smo iz slike razbrali tudi želeno gesto rok s prsti. Z uporabo tehnologije globokega učenja smo iskanje značilnosti prepustili umetni inteligenci, a smo zato na začetku potrebovali veliko bazo že rešenih primerov prepoznav. Dognanja iz tega dela dajejo dobra izhodišča vsem raziskovalcem in inženirjem pri nadaljnjemu raziskovanju in implementaciji sistemov slikovne prepoznave, ki temeljijo na tehnologiji strojnega vida ali globokega učenja.

Keywords

slikovno prepoznavanje;globoko učenje;strojni vid;prsti;magistrske naloge;

Data

Language:	Slovenian
Year of publishing:	2021
Typology:	2.09 - Master's Thesis
Organization:	UM FERI - Faculty of Electrical Engineering and Computer Science
Publisher:	[R. Kopušar]
UDC:	004.932:004.85(043.2)
COBISS:	67960835
Views:	778
Downloads:	56
Average score:	0 (0 votes)
Metadata:

Other data

Secondary language:	English
Secondary title:	Finger recognition using deep learning
Secondary abstract:	The master‘s thesis describes hand fingers recognitions problematic, with which we could in a background manage different tasks and processes. It is designed to provide a solution of same problematic using two different approaches and provide theirs advantages and disadvantages. It directly analyzes this problematic using Pattern image matching technology. We were looking for a pattern in image, with which we could detect and classify hand fingers gesture. Using Deep learning technology, an artificial intelligence has been used to search for a pattern in image, but for this we needed a large image database with solved cases of fingers classification. Findings arising from this thesis give good basis to researchers and engineers to make further development and implementation of image recognitions systems based on Machine visions or Deep learning technology.
Secondary keywords:	image recognition;deep learning;LabVIEW;TensorFlow;finger;
Type (COBISS):	Master's thesis/paper
Thesis comment:	Univ. v Mariboru, Fak. za elektrotehniko, računalništvo in informatiko, Elektrotehnika
Pages:	X, 73 str.
ID:	12913432