kth.sePublications
Change search
CiteExportLink to record
Permanent link

Direct link
Cite
Citation style
  • apa
  • 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
3D Shape Retrieval Meets Machine Learning: Improving Searchability of Industrial 3D Models Using CNN Autoencoder and k-NN
KTH, School of Electrical Engineering and Computer Science (EECS).
2024 (English)Independent thesis Advanced level (degree of Master (Two Years)), 20 credits / 30 HE creditsStudent thesis
Abstract [en]

In the rapidly evolving realm of CAD model retrieval, efficient 3D shape retrieval and reuse remains a challenge in some CAD systems, causing increased design times and costs. This research sought to bridge this gap, investigating the efficacy of machine learning algorithms like 3D CNN Autoencoding and K-NN clustering in enhancing CAD model retrieval. Following an applied research methodology a 3D Shape Retrieval Tool (SRT) prototype was developed. Quantitative evaluation results showed a promising mean Average Precision (mAP) of 69%. Qualitative evaluation results included feedback from three end-users underscoring the need for advanced filtering and seamless tool integration. The findings provide a foundation for future work, highlighting potential areas like dataset augmentation, incorporating natural language processing, or harnessing superior hardware for improved performance.

Abstract [sv]

Inom det snabbt utvecklande området för CAD-modellhämtning, förblir effektiv hämtning och återanvändning av existerande 3D-modeller en utmaning i vissa CAD/PLM-system, vilket leder till ökade designtider och kostnader. Denna forskning syftade till att överbrygga detta glapp genom att undersöka effektiviteten av maskininlärningsalgoritmer som 3D CNN Autoencoding och K-NN klustering för att förbättra CAD-modellhämtning. Med hjälp av en tillämpad forskningsmetodik utvecklades en 3D prototyp. Kvantitativa utvärderingsresultat visade en lovande genomsnittlig precision på 69%. Kvalitativa utvärderingsresultat inkluderade återkoppling från tre slutanvändare som underströk behovet av avancerad filtrering och smidig verktygsintegration. Resultaten ger en grund för framtida arbete, och lyfter fram potentiella områden som dataset augmentation, införande av naturlig språkbearbetning eller användning av avancerad hårdvara för förbättrad prestanda.

Place, publisher, year, edition, pages
2024. , p. 57
Series
TRITA-EECS-EX ; 2024:95
Keywords [en]
3D shape retrieval, pattern recognition, machine learning, autoencoder, k-NN
Keywords [sv]
3D shape retrieval, pattern recognition, machine learning, autoenconder, k-NN
National Category
Computer Sciences
Identifiers
URN: urn:nbn:se:kth:diva-351659OAI: oai:DiVA.org:kth-351659DiVA, id: diva2:1888229
External cooperation
Scania CV AB
Subject / course
Computer Science
Educational program
Master of Science - Computer Science
Supervisors
Examiners
Available from: 2024-08-13 Created: 2024-08-12 Last updated: 2024-08-13Bibliographically approved

Open Access in DiVA

fulltext(2826 kB)447 downloads
File information
File name FULLTEXT01.pdfFile size 2826 kBChecksum SHA-512
43073ff302e270cdc3789e3b67fe19df76678bee29ee8cd24ea40b1468b7b867f09a2ad6465d04062033bd9febfd2a481378db4d820cded9bc6b6cec09fa4d93
Type fulltextMimetype application/pdf

By organisation
School of Electrical Engineering and Computer Science (EECS)
Computer Sciences

Search outside of DiVA

GoogleGoogle Scholar
Total: 448 downloads
The number of downloads is the sum of all downloads of full texts. It may include eg previous versions that are now no longer available

urn-nbn

Altmetric score

urn-nbn
Total: 283 hits
CiteExportLink to record
Permanent link

Direct link
Cite
Citation style
  • apa
  • 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