Uparivanje točaka stereograma Marr-Poggiovom neuronskom mrežom

scientific master's theses

• Author: Marko Bušić
• Mentor: Slobodan Ribarić (mentor)
• Committee member: Leo Budin (predsjednik povjerenstva)
• Committee member: Slobodan Ribarić (član povjerenstva)
• Committee member: Nikola Pavešić (član povjerenstva)
• Granter: University of Zagreb; Faculty of Electrical Engineering and Computing; (Department of Electronics, Microelectronics, Computer and Intelligent Systems); Zagreb
• Defense date and country: 2002, Croatia
• Scientific / art field, discipline and subdiscipline: TECHNICAL SCIENCES; Computing
• Universal decimal classification (UDC): 004 - Computer science and technology. Computing. Data processing

Abstract

Kreiran je programski sustav za rješavanje problema korespondencije u dvookom gledanju trodimenzionalnih scena. Postupak je temeljen na uporabi Marr-Poggiove neuronske mreže. Kreirani sustav ispitan je na stereogramima slučajnih točaka, te na stereogramima stvarnih scena.
Sustav se sastoji od podsustava za kalibraciju i podsustava za uparivanje. Podsustav za kalibraciju određuje faktore rotacije, translacije i skaliranja za transformaciju koordinatnih sustava slika stereograma. Podsustav za uparivanje uparuje lijevu i desnu sliku stereograma. Sustav dobro uparuje stereograme slučajnih točaka i stereograme svijeta blokova, ako povšine imaju teksturu. Zbog prirode algoritma, sustav ne uparuje dobro stereograme stvarnih scena na kojima površine nemaju teksturu.

Abstract (english)

A software system for the stereo correspondence in two-eyed vision of 3D scenes is created. The algorithm is based on the application of Marr-Poggio's neural network. The system is tested on random-dot stereograms and real scene stereograms.
The system consists of the calibration subsystem and of the matching subsystem. The calibration subsystem computes the rotation, translation and scaling factors for the transformation of the coordinate systems of stereogram pictures. The matching subsystem matches the pixels of the left with the pixels of the right stereogram picture.
The system produces good results on random-dot stereograms and blockworld stereograms if the surfaces of the block world objects have texture. Due to the nature of the algorithm, the system does not produce good results if the surfaces do not have texture.

• Keywords: Kooperativni algoritam; dvooko gledanje; stereogram; Marr-Poggiov algoritam; neuronska mreža
• Keywords (english): Cooperative algorithm; stereo vision; stereogram; Marr-Poggio’s algorithm; neural network
• Language: croatian
• URN:NBN: urn:nbn:hr:168:830096
• Study programme: Title: Postgraduate master programe in computing; Study programme type: university; Study level: postgraduate; Academic / professional title: magistar inženjer računarstva (magistar inženjer računarstva)
• Type of resource: Text
• File origin: Born digital
• Access conditions: Closed access
• Created on: 2019-04-24 07:24:42