TY - GEN
T1 - Comparison of objective functions in CNN-based prostate magnetic resonance image segmentation
AU - Mun, Juhyeok
AU - Jang, Won Dong
AU - Sung, Deuk Jae
AU - Kim, Chang-Su
N1 - Funding Information:
This work was supported partly by the National Research Foundation of Korea(NRF) grant funded by the Korea government(MSIP)(No. NRF-2015R1A2A1A10055037), and partly by the Agency for Defense Development (ADD) and Defense Acquisition Program Administration (DAPA) of Korea (UC160016FD).
Publisher Copyright:
© 2017 IEEE.
PY - 2018/2/20
Y1 - 2018/2/20
N2 - We investigate the impacts of objective functions on the performance of deep-learning-based prostate magnetic resonance image segmentation. To this end, we first develop a baseline convolutional neural network (BCNN) for the prostate image segmentation, which consists of encoding, bridge, decoding, and classification modules. In the BCNN, we use 3D convolutional layers to consider volumetric information. Also, we adopt the residual feature forwarding and intermediate feature propagation techniques to make the BCNN reliably trainable for various objective functions. We compare six objective functions: Hamming distance, Euclidean distance, Jaccard index, dice coefficient, cosine similarity, and cross entropy. Experimental results on the PROMISE12 dataset demonstrate that the cosine similarity provides the best segmentation performance, whereas the cross entropy performs the worst.
AB - We investigate the impacts of objective functions on the performance of deep-learning-based prostate magnetic resonance image segmentation. To this end, we first develop a baseline convolutional neural network (BCNN) for the prostate image segmentation, which consists of encoding, bridge, decoding, and classification modules. In the BCNN, we use 3D convolutional layers to consider volumetric information. Also, we adopt the residual feature forwarding and intermediate feature propagation techniques to make the BCNN reliably trainable for various objective functions. We compare six objective functions: Hamming distance, Euclidean distance, Jaccard index, dice coefficient, cosine similarity, and cross entropy. Experimental results on the PROMISE12 dataset demonstrate that the cosine similarity provides the best segmentation performance, whereas the cross entropy performs the worst.
KW - 3D convolutional neural networks
KW - Medical image segmentation
KW - Objective functions
KW - Prostate segmentation
UR - http://www.scopus.com/inward/record.url?scp=85045344961&partnerID=8YFLogxK
U2 - 10.1109/ICIP.2017.8297005
DO - 10.1109/ICIP.2017.8297005
M3 - Conference contribution
AN - SCOPUS:85045344961
T3 - Proceedings - International Conference on Image Processing, ICIP
SP - 3859
EP - 3863
BT - 2017 IEEE International Conference on Image Processing, ICIP 2017 - Proceedings
PB - IEEE Computer Society
T2 - 24th IEEE International Conference on Image Processing, ICIP 2017
Y2 - 17 September 2017 through 20 September 2017
ER -