Dynamic multi-scale CNN forest learning for automatic cervical cancer segmentation

TY - GEN

T1 - Dynamic multi-scale CNN forest learning for automatic cervical cancer segmentation

AU - Bnouni, Nesrine

AU - Rekik, Islem

AU - Rhim, Mohamed Salah

AU - Amara, Najoua Essoukri Ben

PY - 2018

Y1 - 2018

N2 - Deep-learning based labeling methods have gained unprecedented popularity in different computer vision and medical image segmentation tasks. However, to the best of our knowledge, these have not been used for cervical tumor segmentation. More importantly, while the majority of innovative deep-learning works using convolutional neural networks (CNNs) focus on developing more sophisticated and robust architectures (e.g., ResNet, U-Net, GANs), there is very limited work on how to aggregate different CNN architectures to improve their relational learning at multiple levels of CNN-to-CNN interactions. To address this gap, we introduce a Dynamic Multi-Scale CNN Forest (CK+1DMF), which aims to address three major issues in medical image labeling and ensemble CNN learning: (1) heterogeneous distribution of MRI training patches, (2) a bi-directional flow of information between two consecutive CNNs as opposed to cascading CNNs—where information passes in a directional way from current to the next CNN in the cascade, and (3) multiscale anatomical variability across patients. To solve the first issue, we group training samples into K clusters, then design a forest with (K+ 1) trees: a principal tree of CNNs trained using all data samples and subordinate trees, each trained using a cluster of samples. As for the second and third issues, we design each dynamic multiscale tree (DMT) in the forest such that each node in the tree nests a CNN architecture. Two successive CNN nodes in the tree pass bidirectional contextual maps to progressively improve the learning of their relational non-linear mapping. Besides, as we traverse a path from the root node to a leaf node in the tree, the architecture of each CNN node becomes shallower to take in smaller training patches. Our CK+1DMF significantly (p < 0.05) outperformed several conventional and ensemble CNN architectures, including conventional CNN (improvement by 10.3%) and CNN-based DMT (improvement by 5%).

AB - Deep-learning based labeling methods have gained unprecedented popularity in different computer vision and medical image segmentation tasks. However, to the best of our knowledge, these have not been used for cervical tumor segmentation. More importantly, while the majority of innovative deep-learning works using convolutional neural networks (CNNs) focus on developing more sophisticated and robust architectures (e.g., ResNet, U-Net, GANs), there is very limited work on how to aggregate different CNN architectures to improve their relational learning at multiple levels of CNN-to-CNN interactions. To address this gap, we introduce a Dynamic Multi-Scale CNN Forest (CK+1DMF), which aims to address three major issues in medical image labeling and ensemble CNN learning: (1) heterogeneous distribution of MRI training patches, (2) a bi-directional flow of information between two consecutive CNNs as opposed to cascading CNNs—where information passes in a directional way from current to the next CNN in the cascade, and (3) multiscale anatomical variability across patients. To solve the first issue, we group training samples into K clusters, then design a forest with (K+ 1) trees: a principal tree of CNNs trained using all data samples and subordinate trees, each trained using a cluster of samples. As for the second and third issues, we design each dynamic multiscale tree (DMT) in the forest such that each node in the tree nests a CNN architecture. Two successive CNN nodes in the tree pass bidirectional contextual maps to progressively improve the learning of their relational non-linear mapping. Besides, as we traverse a path from the root node to a leaf node in the tree, the architecture of each CNN node becomes shallower to take in smaller training patches. Our CK+1DMF significantly (p < 0.05) outperformed several conventional and ensemble CNN architectures, including conventional CNN (improvement by 10.3%) and CNN-based DMT (improvement by 5%).

UR - https://www.scopus.com/pages/publications/85054476577

U2 - 10.1007/978-3-030-00919-9_3

DO - 10.1007/978-3-030-00919-9_3

M3 - Conference contribution

AN - SCOPUS:85054476577

SN - 9783030009182

VL - 11046

T3 - Lecture Notes in Computer Science

SP - 19

EP - 27

BT - Machine Learning in Medical Imaging

A2 - Liu, Mingxia

A2 - Suk, Heung-Il

A2 - Shi, Yinghuan

PB - Springer Verlag

CY - Switzerland

T2 - 9th International Workshop on Machine Learning in Medical Imaging, MLMI 2018 held in conjunction with the 21st International Conference on Medical Image Computing and Computer-Assisted Intervention, MICCAI 2018

Y2 - 16 September 2018 through 16 September 2018

ER -