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
Denoising Monte Carlo Dose Calculations Using a Deep Neural Network
KTH, School of Electrical Engineering and Computer Science (EECS).
2019 (English)Independent thesis Advanced level (degree of Master (Two Years)), 20 credits / 30 HE creditsStudent thesisAlternative title
Brusreducering av Monte Carlo-dosberäkningar med ett djupt neuralt nätverk (Swedish)
Abstract [en]

This thesis explores the possibility of using a deep neural network (DNN) to denoise Monte Carlo dose calculations for external beam radiotherapy. The dose distributions considered here are for inhomogeneous materials such as those of the human body. The purpose of the project is to explore whether a DNN is able to preserve important features of the dose distributions as well as to evaluate if there is a potential performance gain of using a DNN compared to the traditional approach of running a full Monte Carlo simulation. The network architecture considered in this thesis is a 3D version of the U-net. The results of using the 3D U-net for denoising suggest that it preserves the features of the dose distributions rather well while having a low propagation time. Thus, this indicates that the proposed approach could be a feasible alternative to quickly predict the final dose distribution.

Abstract [sv]

Detta examensarbete utforskar möjligheten att använda ett djupt neuralt nätverk (DNN) för brusreducering av Monte Carlo-dosberäkningar för extern strålbehandling. De dosdistributioner som avses här är för inhomogena material så som människokroppen. Syftet med detta projekt är att avgöra huruvida ett DNN kan bevara dosdistributionernas viktiga attribut och om ett DNN kan öka prestandan jämfört med att beräkna dosen med en komplett Monte Carlosimulering. Nätverksarkitekturen som används i detta projekt är en 3D-version av U-net. Resultaten av att använda ett 3D U-net för avbrusning indikerar att metoden bevarar dosdistributionernas attribut relativt väl och har dessutom låg propageringstid. Detta indikerar alltså att den föreslagna metoden skulle kunna vara ett möjligt alternativ för att snabbt beräkna den slutgiltiga dosen.

Place, publisher, year, edition, pages
2019. , p. 58
Series
TRITA-EECS-EX ; 2019:539
National Category
Computer and Information Sciences
Identifiers
URN: urn:nbn:se:kth:diva-263096OAI: oai:DiVA.org:kth-263096DiVA, id: diva2:1366439
External cooperation
Elekta
Supervisors
Examiners
Available from: 2019-11-18 Created: 2019-10-29 Last updated: 2019-11-18Bibliographically approved

Open Access in DiVA

fulltext(3100 kB)12 downloads
File information
File name FULLTEXT01.pdfFile size 3100 kBChecksum SHA-512
9d71b888f53f1f72a6b34142147e1b56d58385ab2363d72d889ea8b12361aaa73a68e6d9b858a8d32b7afbe5b197026c77092883773bffdfa69284cdff231835
Type fulltextMimetype application/pdf

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

Search outside of DiVA

GoogleGoogle Scholar
Total: 12 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: 17 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