Mid Sweden University

miun.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
Computer-vision as an ErrorRecognition Tool for IndustryProduction Lines: A comparative study of models and neural networks
Mid Sweden University, Faculty of Science, Technology and Media, Department of Computer and Electrical Engineering (2023-).
2023 (English)Independent thesis Advanced level (professional degree), 20 credits / 30 HE creditsStudent thesis
Abstract [en]

To use machine learning as a tool to help humans perform tasks is a common occurrence these days. Machine learning integration in the wood processing industry however, is more rare. Therefore it is suitable to research if machine learning works for some problem in this area. After all, it is unknown which methods are the most suitable for this task. The problem is to recognize when a wooden plank is somehow erroneous within a 30 seconds video clip from a production line. Eight approaches were initially chosen and from these, four was ultimately selected to be compared. The approaches that will be used to investigate the problem are ZFNet, DenseNet, Convolutional autoencoder with SVM, and YOLO. These will be compared head to head, to find which method overall fits this problem best. All methods show good results according to the metrics. What is discovered is that ZFNet is the most likely the best method for this scenario. This is because even though DenseNet and YOLO has a perfect F1 score, the likelihood of overfitting is high based on the graphs and in combination with that, ZFNet has the fastest prediction time, thus ZFNet is concluded to be the best method. From here future work can build upon which methods fits this scenario, for example in a multiclass classification problem which is the immediate follow up. Since ZFNet is a simple model as the classes increase there might be a need to switch to DenseNet to match the complexity better.

Place, publisher, year, edition, pages
2023. , p. 93
Keywords [en]
Computer-vision, Machine Learning, Wood Industry, Neural Network, Convolutional Neural Network, Anomaly Detection, Outlier Detection, Object Detection
National Category
Computer Systems
Identifiers
URN: urn:nbn:se:miun:diva-47907Local ID: DT-H22-A2-004OAI: oai:DiVA.org:miun-47907DiVA, id: diva2:1744461
Subject / course
Computer Engineering DT1
Educational program
Master of Science in Engineering - Computer Engineering TDTEA 300 higher education credits
Supervisors
Examiners
Available from: 2023-03-20 Created: 2023-03-20 Last updated: 2025-09-25Bibliographically approved

Open Access in DiVA

fulltext(2678 kB)357 downloads
File information
File name FULLTEXT01.pdfFile size 2678 kBChecksum SHA-512
25b8cdc0a647ea19bcf338d29177d7851d8ed1f349a9e6614715b2fe681db15357a9d7982b7b14f17d7e9901f602bc0aa81d6bdff342dd069a472a877c30748f
Type fulltextMimetype application/pdf

Search in DiVA

By author/editor
Lundholm, André
By organisation
Department of Computer and Electrical Engineering (2023-)
Computer Systems

Search outside of DiVA

GoogleGoogle Scholar
Total: 361 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: 366 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