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
Finding wooden knots in images using ConvNets
KTH, School of Computer Science and Communication (CSC). (Computer Vision and Active Perception Laboratory)
2015 (English)Independent thesis Advanced level (degree of Master (Two Years)), 20 credits / 30 HE creditsStudent thesisAlternative title
ConvNets för att finna kvisthål i bilder (Swedish)
Abstract [en]

In this thesis ConvNets are used to localize wooden knots in images.The method is compared with a previous method using Kernel SVM with HOG descriptors. The new method is found to work better (F1-score 0.760 vs 0.695). The best performance is achieved by fine-tuning an ImageNet classifier for this domain. The number of negative examples is found to be important, as well as the manner in which mean subtraction is done.

Abstract [sv]

I den här masteruppsatsen används faltande neurala nätverk för att lokalisera kvistar i bilder. Metoden jämförs med en tidigare använd metodbaserad på Kernel SVM med HOG-deskriptorer. Den nya metoden fungerar bättre (F1-score 0.760 vs 0.695). Bäst resultat fås när en klassificerare för ImageNet finjusteras för detta område. Antalet negativa exempel och de exakta detaljerna kring hur medelvärdessubtraktionen sker är viktiga för prestandan.

Place, publisher, year, edition, pages
2015. , p. 60
Keywords [en]
localization, wood, knots, convnets, deep learning
National Category
Computer Vision and Robotics (Autonomous Systems)
Identifiers
URN: urn:nbn:se:kth:diva-210913OAI: oai:DiVA.org:kth-210913DiVA, id: diva2:1121052
External cooperation
Optonova
Subject / course
Computer Science
Educational program
Master of Science in Engineering -Engineering Physics
Presentation
2015-10-08, Stockholm, 16:24 (English)
Supervisors
Examiners
Available from: 2017-10-16 Created: 2017-07-08 Last updated: 2018-01-13Bibliographically approved

Open Access in DiVA

No full text in DiVA

By organisation
School of Computer Science and Communication (CSC)
Computer Vision and Robotics (Autonomous Systems)

Search outside of DiVA

GoogleGoogle Scholar

urn-nbn

Altmetric score

urn-nbn
Total: 57 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