Segmentacija slika dobivenih dubinskom kamerom za detekciju osoba

master's thesis

• Title (english): Image segmentation of range images for person detection
• Author: Josip Šarlija
• Mentor: Sven Lončarić (mentor)
• Committee member: Sven Lončarić (predsjednik povjerenstva)
• Committee member: Marko Subašić (član povjerenstva)
• Committee member: Zvonko Kostanjčar (član povjerenstva)
• Granter: University of Zagreb; Faculty of Electrical Engineering and Computing; Zagreb
• Defense date and country: 2014-07-01, Croatia
• Scientific / art field, discipline and subdiscipline: TECHNICAL SCIENCES; Computing

Abstract

Ovim radom razvijen je sustav za segmentiranje slika dobivenih dubinskom kamerom i generiranje efekata pomoću dobivene segmentacije. Sustav je podijeljen u tri osnovna dijela. Prvi dio sustava je podsustav za komunikaciju s dubinskom kamerom koji upravlja kamerom i dohvaća željene podatke. Drugi dio sustava je podsustav za segmentaciju koji sadrži konkretne implementacije segmentacijskih algoritama. Segmentacijski algoritmi pomoću podataka s dubinske kamere generiraju efekte zamjene pozadine, brisanje korisnika sa slike ili prikaz smanjenog korisnika. Posljednji dio sustava je sam segmentator koji ujedinjava prethodna dva sustava, pruža unificirano sučelje za korištenje sustava i prosljeđuje podatke iz podsustava za komunikaciju sa senzorom u podsustav za segmentaciju. Prikazani su dobiveni rezultati i njihova moguća primjena.

Abstract (english)

Within this paper a system for segmentation of depth images acquired from depth sensors has been developed. Developed system is divided into three separate parts. First part of the system contains a subsystem for communication with the depth sensor. The communication subsystem is responsible for handling cameras resources, initializing and stopping the camera and retrieving data. Second part of the system is the subsystem for segmentation which contains multiple implementations of segmentation algorithms. Segmentation algorithms use depth data from depth sensor to generate the effects of background and player removal or to shrink the player. Last part of the system is the segmentator. Segmentator combines previous subsystems into a whole, provides unified interface for the whole system and transfers data from communication subsystem to segmentation algorithms. Lastly, results of system testing and possible usage are shown.

• Keywords: dubinska kamera; Kinect; segmentacija slike; računalni vid; OpenCV; Kinect SDK
• Keywords (english): depth camera; Kinect; image segmentation; computer vision; OpenCV
• Language: croatian
• URN:NBN: urn:nbn:hr:168:854926
• Study programme: Title: Computing; Study programme type: university; Study level: graduate; Academic / professional title: magistar/magistra inženjer/inženjerka računarstva (magistar/magistra inženjer/inženjerka računarstva)
• Type of resource: Text
• File origin: Born digital
• Access conditions: Closed access
• Created on: 2018-02-06 12:13:23