Högskolan i Skövde

his.sePublications
Change search
CiteExportLink to record
Permanent link

Direct link
Cite
Citation style
  • apa
  • apa-cv
  • 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
Diagnosis of lumbar spinal stenosis using deep learning of MRIs
University of Skövde, School of Bioscience.
2024 (English)Independent thesis Advanced level (degree of Master (One Year)), 20 credits / 30 HE creditsStudent thesis
Abstract [en]

This thesis examines the potential of Deep Learning (DL) technologies in improving the diagnosis of Lumbar Spinal Stenosis (LSS) through MRI analyses. Current diagnostic methods for LSS rely heavily on the subjective interpretations of radiologists, leading to inconsistencies and potential misdiagnoses. This study aimed to develop a DL-based object detection model to minimize these issues, enhancing both the accuracy and efficiency of MRI evaluations. The research employed a pre-trained YOLO (You Only Look Once) model, adapted through transfer learning to detect signs of canal stenosis in lumbar spine MRI images. The open-source datasets of lumbar spine MRIs were preprocessed and used for training and validating the model, employing cross-validation techniques to optimize and assess its performance. The findings revealed that the adapted YOLO model achieved a mean Average Precision (mAP) of 97.1% and an Intersection Over Union (IOU) threshold of 0.5, with a classification accuracy of 93%. While these results are impressive, it is important to note that a direct comparison with existing diagnostic methods is not feasible due to differences in datasets. The dataset used in this study lacks documented labeling criteria, unlike the balanced and reliable private datasets used in prior studies. The model demonstrated high scores across various metrics, including a precision of 91.9%, recall of 93.5%, and an F1-score of 93%. This study highlights the prominent role of transfer learning in achieving high performance, suggesting that the implementation of this DL model could significantly reduce human errors and the workload on radiologists, accelerate diagnostic processes, and integrate seamlessly into clinical settings. However, limitations such as class imbalance in training data suggest the need for further research using balanced, comprehensive datasets to fully harness the capabilities of DL in clinical diagnostics. Page 

Place, publisher, year, edition, pages
2024. , p. 49
National Category
Bioinformatics and Computational Biology
Identifiers
URN: urn:nbn:se:his:diva-24580OAI: oai:DiVA.org:his-24580DiVA, id: diva2:1901273
Subject / course
Bioinformatics
Educational program
Bioinformatics - Master’s Programme
Supervisors
Examiners
Available from: 2024-09-26 Created: 2024-09-26 Last updated: 2025-09-29Bibliographically approved

Open Access in DiVA

fulltext(1086 kB)211 downloads
File information
File name FULLTEXT01.pdfFile size 1086 kBChecksum SHA-512
3254ee1a179b9e327a65883b88258d69e0554285f6c1b4cd469bef4620bc362b2b1f1d67142e3e6d583afb032c6a5f5c6423dba1cb17818bc7bcc12c6bad0352
Type fulltextMimetype application/pdf

By organisation
School of Bioscience
Bioinformatics and Computational Biology

Search outside of DiVA

GoogleGoogle Scholar
Total: 211 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: 333 hits
CiteExportLink to record
Permanent link

Direct link
Cite
Citation style
  • apa
  • apa-cv
  • 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