Automatic 3D ultrasound segmentation of the first trimester placenta using deep learning
Placental volume measured with 3D ultrasound in the first trimester has been shown to be correlated to adverse pregnancy outcomes. This could potentially be used as a screening test to predict the 'at risk' pregnancy. However, manual segmentation whilst previously shown to be accurate and...
| Main Authors: | , , , , , , |
|---|---|
| Format: | Conference item |
| Published: |
Institute of Electrical and Electronics Engineers
2017
|
| _version_ | 1797070528108822528 |
|---|---|
| author | Looney, P Stevenson, G Nicolaides, K Plasencia, W Molloholli, M Natsis, S Collins, S |
| author_facet | Looney, P Stevenson, G Nicolaides, K Plasencia, W Molloholli, M Natsis, S Collins, S |
| author_sort | Looney, P |
| collection | OXFORD |
| description | Placental volume measured with 3D ultrasound in the first trimester has been shown to be correlated to adverse pregnancy outcomes. This could potentially be used as a screening test to predict the 'at risk' pregnancy. However, manual segmentation whilst previously shown to be accurate and repeatable is very time consuming and semi-automated methods still require operator input. To generate a screening tool, fully automated placental segmentation is required. In this work, a deep convolutional neural network (cNN), DeepMedic, was trained using the output of the semi-automated Random Walker method as ground truth. 300 3D ultrasound scans of first trimester placentas were used to train, validate and test the cNN. Compared against the semi-automated segmentation, resultant median (1 st Quartile, 3 rd Quartile) Dice Similarity Coefficient was 0.73 (0.66, 0.76). The median (1 st Quartile, 3 rd Quartile) Hausdorff distance was 27 mm (18 mm, 36 mm). We present the first attempt at using a deep cNN for segmentation of 3D ultrasound of the placenta. This work shows that feasible results compared to ground truth were obtained that could form the basis of a fully automatic segmentation method. |
| first_indexed | 2024-03-06T22:40:10Z |
| format | Conference item |
| id | oxford-uuid:5b478d5d-37f8-4601-8a7c-f3aa48c52393 |
| institution | University of Oxford |
| last_indexed | 2024-03-06T22:40:10Z |
| publishDate | 2017 |
| publisher | Institute of Electrical and Electronics Engineers |
| record_format | dspace |
| spelling | oxford-uuid:5b478d5d-37f8-4601-8a7c-f3aa48c523932022-03-26T17:21:05ZAutomatic 3D ultrasound segmentation of the first trimester placenta using deep learningConference itemhttp://purl.org/coar/resource_type/c_5794uuid:5b478d5d-37f8-4601-8a7c-f3aa48c52393Symplectic Elements at OxfordInstitute of Electrical and Electronics Engineers2017Looney, PStevenson, GNicolaides, KPlasencia, WMolloholli, MNatsis, SCollins, SPlacental volume measured with 3D ultrasound in the first trimester has been shown to be correlated to adverse pregnancy outcomes. This could potentially be used as a screening test to predict the 'at risk' pregnancy. However, manual segmentation whilst previously shown to be accurate and repeatable is very time consuming and semi-automated methods still require operator input. To generate a screening tool, fully automated placental segmentation is required. In this work, a deep convolutional neural network (cNN), DeepMedic, was trained using the output of the semi-automated Random Walker method as ground truth. 300 3D ultrasound scans of first trimester placentas were used to train, validate and test the cNN. Compared against the semi-automated segmentation, resultant median (1 st Quartile, 3 rd Quartile) Dice Similarity Coefficient was 0.73 (0.66, 0.76). The median (1 st Quartile, 3 rd Quartile) Hausdorff distance was 27 mm (18 mm, 36 mm). We present the first attempt at using a deep cNN for segmentation of 3D ultrasound of the placenta. This work shows that feasible results compared to ground truth were obtained that could form the basis of a fully automatic segmentation method. |
| spellingShingle | Looney, P Stevenson, G Nicolaides, K Plasencia, W Molloholli, M Natsis, S Collins, S Automatic 3D ultrasound segmentation of the first trimester placenta using deep learning |
| title | Automatic 3D ultrasound segmentation of the first trimester placenta using deep learning |
| title_full | Automatic 3D ultrasound segmentation of the first trimester placenta using deep learning |
| title_fullStr | Automatic 3D ultrasound segmentation of the first trimester placenta using deep learning |
| title_full_unstemmed | Automatic 3D ultrasound segmentation of the first trimester placenta using deep learning |
| title_short | Automatic 3D ultrasound segmentation of the first trimester placenta using deep learning |
| title_sort | automatic 3d ultrasound segmentation of the first trimester placenta using deep learning |
| work_keys_str_mv | AT looneyp automatic3dultrasoundsegmentationofthefirsttrimesterplacentausingdeeplearning AT stevensong automatic3dultrasoundsegmentationofthefirsttrimesterplacentausingdeeplearning AT nicolaidesk automatic3dultrasoundsegmentationofthefirsttrimesterplacentausingdeeplearning AT plasenciaw automatic3dultrasoundsegmentationofthefirsttrimesterplacentausingdeeplearning AT mollohollim automatic3dultrasoundsegmentationofthefirsttrimesterplacentausingdeeplearning AT natsiss automatic3dultrasoundsegmentationofthefirsttrimesterplacentausingdeeplearning AT collinss automatic3dultrasoundsegmentationofthefirsttrimesterplacentausingdeeplearning |