TY - GEN
T1 - Boosting hand-crafted features for curvilinear structure segmentation by learning context filters
AU - Annunziata, Roberto
AU - Kheirkhah, Ahmad
AU - Hamrah, Pedram
AU - Trucco, Emanuele
N1 - Funding Information:
This research was supported by the EU Marie Curie ITN 316990 “REVAMMAD”. The authors are grateful to Amos Sironi (CVlab, EPFL) for providing the VC6 and BF2D datasets and to Max Law (Western University) for the OOF code.
Publisher Copyright:
© Springer International Publishing Switzerland 2015.
PY - 2015
Y1 - 2015
N2 - Combining hand-crafted features and learned filters (i.e. feature boosting) for curvilinear structure segmentation has been proposed recently to capture key structure configurations while limiting the number of learned filters. Here, we present a novel combination method pairing hand-crafted appearance features with learned context filters. Unlike recent solutions based only on appearance filters, our method introduces context information in the filter learning process. Moreover, it reduces the potential redundancy of learned appearance filters that may be reconstructed using a combination of hand-crafted filters. Finally, the use of k-means for filter learning makes it fast and easily adaptable to other datasets, even when large dictionary sizes (e.g. 200 filters) are needed to improve performance. Comprehensive experimental results using 3 challenging datasets show that our combination method outperforms recent state-of-the-art HCFs and a recent combination approach for both performance and computational time.
AB - Combining hand-crafted features and learned filters (i.e. feature boosting) for curvilinear structure segmentation has been proposed recently to capture key structure configurations while limiting the number of learned filters. Here, we present a novel combination method pairing hand-crafted appearance features with learned context filters. Unlike recent solutions based only on appearance filters, our method introduces context information in the filter learning process. Moreover, it reduces the potential redundancy of learned appearance filters that may be reconstructed using a combination of hand-crafted filters. Finally, the use of k-means for filter learning makes it fast and easily adaptable to other datasets, even when large dictionary sizes (e.g. 200 filters) are needed to improve performance. Comprehensive experimental results using 3 challenging datasets show that our combination method outperforms recent state-of-the-art HCFs and a recent combination approach for both performance and computational time.
UR - http://www.scopus.com/inward/record.url?scp=84951746253&partnerID=8YFLogxK
UR - http://www.scopus.com/inward/citedby.url?scp=84951746253&partnerID=8YFLogxK
U2 - 10.1007/978-3-319-24574-4_71
DO - 10.1007/978-3-319-24574-4_71
M3 - Conference contribution
AN - SCOPUS:84951746253
SN - 9783319245737
T3 - Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)
SP - 596
EP - 603
BT - Medical Image Computing and Computer-Assisted Intervention – MICCAI 2015 - 18th International Conference, Proceedings
A2 - Frangi, Alejandro F.
A2 - Navab, Nassir
A2 - Hornegger, Joachim
A2 - Wells, William M.
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 -