Raspoznavanje gesta na temelju 2D slike i na temelju RGBD prikaza

master's thesis

• Title (english): Gesture Recognition based on 2D Images and RGBD Representation
• Author: Kruno Hrvatinić
• Mentor: Zdenko Kovačić (mentor)
• Committee member: Zdenko Kovačić (predsjednik povjerenstva)
• Committee member: Stjepan Bogdan (član povjerenstva)
• Committee member: Ivana Palunko (član povjerenstva)
• Granter: University of Zagreb; Faculty of Electrical Engineering and Computing; Zagreb
• Defense date and country: 2014-07-16, Croatia
• Scientific / art field, discipline and subdiscipline: TECHNICAL SCIENCES; Electrical Engineering

Abstract

U okviru robotom potpomognutog protokola za dijagnozu autizma razvijen je modul za detekciju i praćenje proizvoljnog broja predmeta čije su boje poznate te prepoznavanje gesta u pokretima koje se tim predmetima izvode. Modul je prilagođen za izvođenje na humanoidnom robotu NAO u realnom vremenu. Praćenje predmeta se izvodi na robusan način koji dopušta snažnu interakciju između predmeta slične boje bez da se time gubi identitet pojedinog predmeta, te se pokazalo pouzdanim u realnim uvjetima izvođenja zadatka i u testovima praćenja. Prepoznavanje geste s korisničke je strane intuitivno i kvalitetno prepoznaje pokrete koji se koriste pri izvođenju protokola za dijagnozu.

Abstract (english)

As part of a robot-assisted autism diagnostic protocol, a module has been developed for the detection and tracking of an arbitrary number of objects of known color, as well as recognition of gestures based on the movements made by the objects. The module is built for real-time execution on the NAO humanoid robot. Object tracking is done using a robust method allowing for strong interaction between objects of a similar color without disrupting the tracking of any interacting object, and is shown to be reliable both in real-world conditions as well as in tracking tests. The module's gesture recognition capabilities are intuitive to use and provide quality recognition of all movements used in the diagnostic protocol.

• Keywords: autizam; NAO; humanoidni robot; segmentacija; projekcija histograma; histerezni prag; praćenje; uspoređivanje trajektorije
• Keywords (english): autism; NAO; humanoid robot; segmentation; histogram backprojection; hysteresis thresholding; tracking; trajectory matching
• Language: croatian
• URN:NBN: urn:nbn:hr:168:985066
• Study programme: Title: Electrical Engineering and Information Technology; Study programme type: university; Study level: graduate; Academic / professional title: magistar/magistra inženjer/inženjerka elektrotehnike i informacijske tehnologije (magistar/magistra inženjer/inženjerka elektrotehnike i informacijske tehnologije)
• Type of resource: Text
• File origin: Born digital
• Access conditions: Closed access
• Created on: 2018-02-06 09:12:49