Forskning ved Københavns Universitet - Københavns Universitet

Forside

Accurate segmentation of dental panoramic radiographs with u-NETS

Publikation: Bidrag til bog/antologi/rapportKonferencebidrag i proceedingsForskningfagfællebedømt

Fully convolutional neural networks (FCNs) have proven to be powerful tools for medical image segmentation. We apply an FCN based on the U-Net architecture for the challenging task of semantic segmentation of dental panoramic radiographs and discuss general tricks for improving segmentation performance. Among those are network ensembling, test-time augmentation, data symmetry exploitation and bootstrapping of low quality annotations. The performance of our approach was tested on a highly variable dataset of 1500 dental panoramic radiographs. A single network reached the Dice score of 0.934 where 1201 images were used for training, forming an ensemble increased the score to 0.936.

OriginalsprogEngelsk
TitelISBI 2019 - 2019 IEEE International Symposium on Biomedical Imaging
ForlagIEEE
Publikationsdato2019
Sider15-19
ISBN (Elektronisk)9781538636411
DOI
StatusUdgivet - 2019
Begivenhed16th IEEE International Symposium on Biomedical Imaging, ISBI 2019 - Venice, Italien
Varighed: 8 apr. 201911 apr. 2019

Konference

Konference16th IEEE International Symposium on Biomedical Imaging, ISBI 2019
LandItalien
ByVenice
Periode08/04/201911/04/2019
Sponsoret al., IEEE Engineering in Medicine and Biology Society (EMB), IEEE Signal Processing Society, The Institute of Electrical and Electronics Engineers (IEEE), UAI, United Imaging Intelligence

ID: 230481516