Advanced

Using deep learning to remove computed tomography artifacts due to hip replacement

Zahirovic, Adelina LU (2020) In Master's Theses in Mathematical Sciences FMAM05 20201
Mathematics (Faculty of Engineering)
Abstract
Computed Tomography (CT) is the most common diagnostic method for cancer and prostate cancer is the most common cancer among men in Sweden. Some of the patients that get scanned have hip replacements that cause artifacts in CT scans that make the CT images unreadable for physicians. In this Master’s Thesis an autoencoder model was implemented to reduce artifacts due to hip replacements in CT images. The model was trained, validated and tested on CT images provided by EXINI Diagnostics AB, Lund, Sweden.
The autoencoder was implemented using the deep learning framework Keras in Python. Autoencoders have been used to reduce different types of noise in other experiments and shown great results. The produced results show that the model is a... (More)
Computed Tomography (CT) is the most common diagnostic method for cancer and prostate cancer is the most common cancer among men in Sweden. Some of the patients that get scanned have hip replacements that cause artifacts in CT scans that make the CT images unreadable for physicians. In this Master’s Thesis an autoencoder model was implemented to reduce artifacts due to hip replacements in CT images. The model was trained, validated and tested on CT images provided by EXINI Diagnostics AB, Lund, Sweden.
The autoencoder was implemented using the deep learning framework Keras in Python. Autoencoders have been used to reduce different types of noise in other experiments and shown great results. The produced results show that the model is a good start for further work to completely reduce artifacts due to hip replacement. (Less)
Please use this url to cite or link to this publication:
author
Zahirovic, Adelina LU
supervisor
organization
alternative title
Djupa faltningsnätverk för borttagning av protesorsakade artefakter i datortomografiska bilder
course
FMAM05 20201
year
type
H2 - Master's Degree (Two Years)
subject
keywords
Image analysis
publication/series
Master's Theses in Mathematical Sciences
report number
LUTFMA-3406-2020
ISSN
1404-6342
other publication id
2020:E23
language
English
id
9008353
date added to LUP
2020-05-25 17:00:32
date last changed
2020-05-25 17:00:32
@misc{9008353,
  abstract     = {Computed Tomography (CT) is the most common diagnostic method for cancer and prostate cancer is the most common cancer among men in Sweden. Some of the patients that get scanned have hip replacements that cause artifacts in CT scans that make the CT images unreadable for physicians. In this Master’s Thesis an autoencoder model was implemented to reduce artifacts due to hip replacements in CT images. The model was trained, validated and tested on CT images provided by EXINI Diagnostics AB, Lund, Sweden.
The autoencoder was implemented using the deep learning framework Keras in Python. Autoencoders have been used to reduce different types of noise in other experiments and shown great results. The produced results show that the model is a good start for further work to completely reduce artifacts due to hip replacement.},
  author       = {Zahirovic, Adelina},
  issn         = {1404-6342},
  keyword      = {Image analysis},
  language     = {eng},
  note         = {Student Paper},
  series       = {Master's Theses in Mathematical Sciences},
  title        = {Using deep learning to remove computed tomography artifacts due to hip replacement},
  year         = {2020},
}