Change search
CiteExportLink to record
Permanent link

Direct link
Cite
Citation style
  • apa
  • harvard1
  • ieee
  • modern-language-association-8th-edition
  • vancouver
  • Other style
More styles
Language
  • de-DE
  • en-GB
  • en-US
  • fi-FI
  • nn-NO
  • nn-NB
  • sv-SE
  • Other locale
More languages
Output format
  • html
  • text
  • asciidoc
  • rtf
A Generative Model of Indoor Scenes.
KTH, School of Computer Science and Communication (CSC).
2011 (English)Independent thesis Advanced level (professional degree), 20 credits / 30 HE creditsStudent thesis
Abstract [en]

In this Master thesis project, we tackle the problem of 3D indoor scene understanding from a single image. Following existing approaches, we parameterize the problem as one of estimating the faces of a 3D cuboid. Towards this goal, we propose a generative model of the scene, which takes advantage of sophisticated image features while utilizing a simple prior learned from the statistics of room photographs. We utilize an adaptive Metropolis-Hastings sampling scheme for inference and demonstrate that our approach outperforms the state-of-the-art on the main two benchmarks that exist for this task. Moreover, our approach outperforms previously published generative models for this problem by more than 10% absolute error.

Abstract [sv]

I det har examensarbetet undersöker vi problemet med att förstå inomhusscener i 3D från en monokulär bild. Vi använder oss av existerande ansatser och parametriserar så att problemet reduceras till att finna planen i en konvex polyeder. Vi förslår en generativ modell som använder sig av sofistikerade bildegenskaper tillsammans med en enkel prior-fördelning baserad på statistik från scenbilderna och en adaptiv Metropolis-Hasting samplingstrategi för att finna lösningar. Vi visar att vårt tillvägagångssätt presterar bättre än de nuvarande bästa metoderna på de två bildsamlingar som används som måttstock för detta problem. Utöver det presterar vår modell mer än 10% bättre i absolutfel än föregående generativa modeller.

Place, publisher, year, edition, pages
2011.
Series
Trita-CSC-E, ISSN 1653-5715 ; 2011:138
National Category
Computer Science
Identifiers
URN: urn:nbn:se:kth:diva-130710OAI: oai:DiVA.org:kth-130710DiVA: diva2:654157
Educational program
Master of Science in Engineering -Engineering Physics
Uppsok
Technology
Supervisors
Examiners
Available from: 2013-10-07 Created: 2013-10-07

Open Access in DiVA

No full text

Other links

http://www.nada.kth.se/utbildning/grukth/exjobb/rapportlistor/2011/rapporter11/hjelm_martin_11138.pdf
By organisation
School of Computer Science and Communication (CSC)
Computer Science

Search outside of DiVA

GoogleGoogle Scholar

urn-nbn

Altmetric score

urn-nbn
Total: 22 hits
CiteExportLink to record
Permanent link

Direct link
Cite
Citation style
  • apa
  • harvard1
  • ieee
  • modern-language-association-8th-edition
  • vancouver
  • Other style
More styles
Language
  • de-DE
  • en-GB
  • en-US
  • fi-FI
  • nn-NO
  • nn-NB
  • sv-SE
  • Other locale
More languages
Output format
  • html
  • text
  • asciidoc
  • rtf