Supervoxel based weakly-supervised multi-level 3D CNNs for lung nodule detection and segmentation

Pulmonary nodule detection and segmentation are two important works for early diagnosis and treatment of lung cancer. The work of detection is to locate pulmonary nodules in a given chest CT scan, and the segmentation aims at extracting all the voxels from a CT scan within each nodule’s space. This paper propose a novel framework to process both nodule detection and segmentation integrately, which is implemented as the combination of SLIC supervoxel segmentation and CNN classification. The learning of CNN just require weakly labeled data annotations, where only a single coordinate is provided for each annotated nodule as the ground truth. The CNN architecture is designed as a 3D multi-level framework, which is able to comprehensively recognize nodules with variant sizes and shapes. Experiments on the dataset of LUNA16 challenge expressed prominent detecting performance, demonstrating the necessity and efficiency of 3D CNN architecture and multi-level framework for computer-aided detection of pulmonary nodules. Meanwhile the evaluation of segmentation presented impressive performance, producing elegant shapes of real nodules, which proves the great efficiency of SLIC technique.

[1]  Hengyong Yu,et al.  Deep Learning for the Classification of Lung Nodules , 2016, ArXiv.

[2]  Ling Shao,et al.  Learning Deep and Wide: A Spectral Method for Learning Deep Networks , 2014, IEEE Transactions on Neural Networks and Learning Systems.

[3]  Thomas Brox,et al.  U-Net: Convolutional Networks for Biomedical Image Segmentation , 2015, MICCAI.

[4]  Temesguen Messay,et al.  A new computationally efficient CAD system for pulmonary nodule detection in CT imagery , 2010, Medical Image Anal..

[5]  Hyojin Kim,et al.  Lung nodule detection using 3D convolutional neural networks trained on weakly labeled data , 2016, SPIE Medical Imaging.

[6]  Nima Tajbakhsh,et al.  Convolutional Neural Networks for Medical Image Analysis: Full Training or Fine Tuning? , 2016, IEEE Transactions on Medical Imaging.

[7]  Christian Igel,et al.  Deep Feature Learning for Knee Cartilage Segmentation Using a Triplanar Convolutional Neural Network , 2013, MICCAI.

[8]  Hao Chen,et al.  Multilevel Contextual 3-D CNNs for False Positive Reduction in Pulmonary Nodule Detection , 2017, IEEE Transactions on Biomedical Engineering.

[9]  Bram van Ginneken,et al.  A large-scale evaluation of automatic pulmonary nodule detection in chest CT using local image features and k-nearest-neighbour classification , 2009, Medical Image Anal..

[10]  Wenjing Tang,et al.  An image segmentation algorithm based on improved multiscale random field model in wavelet domain , 2016, J. Ambient Intell. Humaniz. Comput..

[11]  Ling Shao,et al.  Deep Dynamic Neural Networks for Multimodal Gesture Segmentation and Recognition , 2016, IEEE Transactions on Pattern Analysis and Machine Intelligence.

[12]  Max A. Viergever,et al.  On Combining Computer-Aided Detection Systems , 2011, IEEE Transactions on Medical Imaging.

[13]  Sergey Ioffe,et al.  Batch Normalization: Accelerating Deep Network Training by Reducing Internal Covariate Shift , 2015, ICML.

[14]  Nasir M. Rajpoot,et al.  Locality Sensitive Deep Learning for Detection and Classification of Nuclei in Routine Colon Cancer Histology Images , 2016, IEEE Trans. Medical Imaging.

[15]  Bram van Ginneken,et al.  Automatic detection of subsolid pulmonary nodules in thoracic computed tomography images , 2014, Medical Image Anal..

[16]  Kaiming He,et al.  Focal Loss for Dense Object Detection , 2017, 2017 IEEE International Conference on Computer Vision (ICCV).

[17]  Manfred Huber,et al.  Using deep learning to enhance cancer diagnosis and classication , 2013 .

[18]  Nitish Srivastava,et al.  Improving neural networks by preventing co-adaptation of feature detectors , 2012, ArXiv.

[19]  Fang Hao,et al.  Medical image denoising by parallel non-local means , 2016, Neurocomputing.

[20]  Elsa D. Angelini,et al.  Discriminative Localization in CNNs for Weakly-Supervised Segmentation of Pulmonary Nodules , 2017, MICCAI.

[21]  Bram van Ginneken,et al.  Automatic detection of large pulmonary solid nodules in thoracic CT images. , 2015, Medical physics.

[22]  Bram van Ginneken,et al.  Pulmonary Nodule Detection in CT Images: False Positive Reduction Using Multi-View Convolutional Networks , 2016, IEEE Transactions on Medical Imaging.

[23]  Bram van Ginneken,et al.  Off-the-shelf convolutional neural network features for pulmonary nodule detection in computed tomography scans , 2015, 2015 IEEE 12th International Symposium on Biomedical Imaging (ISBI).

[24]  Adin Ramirez Rivera,et al.  Large residual multiple view 3D CNN for false positive reduction in pulmonary nodule detection , 2017, 2017 IEEE Conference on Computational Intelligence in Bioinformatics and Computational Biology (CIBCB).

[25]  Pascal Fua,et al.  SLIC Superpixels Compared to State-of-the-Art Superpixel Methods , 2012, IEEE Transactions on Pattern Analysis and Machine Intelligence.

[26]  Heinz-Otto Peitgen,et al.  Morphological segmentation and partial volume analysis for volumetry of solid pulmonary lesions in thoracic CT scans , 2006, IEEE Transactions on Medical Imaging.