magistrsko delo
Jože Baligač (Author), Božidar Potočnik (Mentor)

Abstract

V magistrskem delu smo se ukvarjali s preučevanjem postopkov za razpoznavo kretenj rok ter razvojem naprednega, naravnega uporabniškega vmesnika. Produkta raziskovanja sta algoritem, zmožen razpoznave desetih različnih statičnih in dinamičnih kretenj, ter napredna aplikacija kot intuitivni vmesnik za interakcijo z računalnikom. Algoritem razpoznave kretenj je zgrajen iz dveh večjih modulov. Prvi je namenjen razpoznavi statičnih kretenj in sestoji iz segmentacije na osnovi barve kože, ekstrakcije Hu-jevih invariantnih momentov in klasifikacije z metodo podpornih vektorjev. Drugi modul, namenjen razpoznavi dinamičnih kretenj, sestoji iz prilagojenega algoritma CAMSHIFT, ki smo ga izboljšali z metodo odštevanja ozadja in preproste klasifikacijske metode, ki za svoje delovanje uporablja krožno vrsto. Razpoznavo statičnih kretenj smo validirali na množici 300 digitalnih slik, kjer je bila največja dosežena uspešnost 86,66 %. Razpoznavo dinamičnih pa na 20 videoposnetkih, kjer smo dosegli 95 % uspešnost. Uporabnost, zadovoljstvo in enostavnost uporabe predlaganega uporabniškega vmesnika smo ocenili s pomočjo eksperimenta in ankete. Rezultati so pokazali, da je predlagan uporabniški vmesnik uporaben, enostaven in hkrati zagotavlja dobro uporabniško izkušnjo ter ga lahko klasificiramo v skupino naravnih uporabniških vmesnikov.

Keywords

obdelava digitalnih slik;računalniški vid;naravni uporabniški vmesnik;razpoznava kretenj;

Data

Language: Slovenian
Year of publishing:
Typology: 2.09 - Master's Thesis
Organization: UM FERI - Faculty of Electrical Engineering and Computer Science
Publisher: [J. Baligač]
UDC: 004.5:004.932(043.2)
COBISS: 18005782 Link will open in a new window
Views: 1572
Downloads: 148
Average score: 0 (0 votes)
Metadata: JSON JSON-RDF JSON-LD TURTLE N-TRIPLES XML RDFA MICRODATA DC-XML DC-RDF RDF

Other data

Secondary language: English
Secondary title: ADVANCED USER INTERFACE BASED ON COMPUTER IMAGE PROCESSING TECHNIQUES
Secondary abstract: In this master's thesis we have dealt with the study of hand gesture recognition techniques and development of advanced, natural user interface. The products of the research is an algorithm, capable of recognizing ten different static and dynamic gestures and an advanced application, sample of intuitive interface for computer interaction. The recognition algorithm has two major modules. The first one is used to detect static gestures and consists of skin color based segmentation, extraction of Hu invariant moments and a Support Vector Machines classifier. The second module is used to detect dynamic gestures and consists of adapted CAMSHIFT algorithm, which was enhanced with background subtraction method. The classification of dynamic gestures was done by simple approach using circular queue. Recognition of static gestures was validated on a set of 300 digital images, with 86.66% efficiency. Recognition of dynamic gesture was validated on set of 20 digital videos, with 95% efficiency. Usability, satisfaction and ease of use of the proposed user interface were validated by simple experiment and a questionnaire. The results showed that proposed user interface is useful, simple and at the same time ensures a good user experience. Certainly, it can be classified into a group of natural user interfaces.
Secondary keywords: digital images processing;computer vision;natural user interface;gesture recognition;
URN: URN:SI:UM:
Type (COBISS): Master's thesis/paper
Thesis comment: Univ. v Mariboru, Fak. za elektrotehniko, računalništvo in informatiko
Pages: IX, 58 f.
ID: 8729850