TY - GEN
T1 - Robust segmentation of various anatomies in 3D ultrasound using hough forests and learned data representations
AU - Milletari, Fausto
AU - Ahmadi, Seyed Ahmad
AU - Kroll, Christine
AU - Hennersperger, Christoph
AU - Tombari, Federico
AU - Shah, Amit
AU - Plate, Annika
AU - Boetzel, Kai
AU - Navab, Nassir
N1 - Publisher Copyright:
© Springer International Publishing Switzerland 2015.
PY - 2015
Y1 - 2015
N2 - 3D ultrasound segmentation is a challenging task due to image artefacts, low signal-to-noise ratio and lack of contrast at anatomical boundaries. Current solutions usually rely on complex, anatomy-specific regularization methods to improve segmentation accuracy. In this work, we propose a highly adaptive learning-based method for fully automatic segmentation of ultrasound volumes. During training, anatomy-specific features are obtained through a sparse auto-encoder. The extracted features are employed in a Hough Forest based framework to retrieve the position of the target anatomy and its segmentation contour. The resulting method is fully automatic, i.e. it does not require any human interaction, and can robustly and automatically adapt to different anatomies yet enforcing appearance and shape constraints.We demonstrate the performance of the method for three different applications: segmentation of midbrain, left ventricle of the heart and prostate.
AB - 3D ultrasound segmentation is a challenging task due to image artefacts, low signal-to-noise ratio and lack of contrast at anatomical boundaries. Current solutions usually rely on complex, anatomy-specific regularization methods to improve segmentation accuracy. In this work, we propose a highly adaptive learning-based method for fully automatic segmentation of ultrasound volumes. During training, anatomy-specific features are obtained through a sparse auto-encoder. The extracted features are employed in a Hough Forest based framework to retrieve the position of the target anatomy and its segmentation contour. The resulting method is fully automatic, i.e. it does not require any human interaction, and can robustly and automatically adapt to different anatomies yet enforcing appearance and shape constraints.We demonstrate the performance of the method for three different applications: segmentation of midbrain, left ventricle of the heart and prostate.
UR - http://www.scopus.com/inward/record.url?scp=84950986731&partnerID=8YFLogxK
U2 - 10.1007/978-3-319-24571-3_14
DO - 10.1007/978-3-319-24571-3_14
M3 - Conference contribution
AN - SCOPUS:84950986731
SN - 9783319245706
SN - 9783319245706
SN - 9783319245706
T3 - Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)
SP - 111
EP - 118
BT - Medical Image Computing and Computer-Assisted Intervention - MICCAI 2015 - 18th International Conference, Proceedings
A2 - Hornegger, Joachim
A2 - Frangi, Alejandro F.
A2 - Wells, William M.
A2 - Frangi, Alejandro F.
A2 - Navab, Nassir
A2 - Hornegger, Joachim
A2 - Navab, Nassir
A2 - Wells, William M.
A2 - Wells, William M.
A2 - Frangi, Alejandro F.
A2 - Hornegger, Joachim
A2 - Navab, Nassir
PB - Springer Verlag
T2 - 18th International Conference on Medical Image Computing and Computer-Assisted Intervention, MICCAI 2015
Y2 - 5 October 2015 through 9 October 2015
ER -